fix(corecmd): support strongly-typed DTOs in event/auto wait modes

waitResource previously asserted result.Data() to map[string]any, which
caused event mode and auto mode to fail with 'data is not an object'
when commands returned strongly-typed DTOs (struct or struct pointer).

Now normalizes via JSON round-trip for non-map types, preserving the
fast path for map[string]any. Added regression tests covering struct
values, struct pointers, nested dotted queries, and error cases.

.changes/released/1.0.59-beta.2/1009-clitrack-telemetry.md

@@ -0,0 +1,5 @@
+---
+category: Added
+---
+
+- **Privacy-safe CLI telemetry** (#1009) — reports reviewed command outcomes and profile identity dimensions while excluding command arguments, output, paths, device fingerprints, and automatic system dimensions; `DO_NOT_TRACK=1` disables reporting.

.changes/released/1.0.59-beta.2/1019-help-feedback-entry.md

@@ -0,0 +1,5 @@
+---
+category: Added
+---
+
+- **Feedback survey entry in root help** (#1019) — `dws --help` now closes with a Feedback section linking the user-experience survey form.

.changes/released/1.0.59-beta.2/chat-reference-card-verification.md

@@ -0,0 +1,6 @@
+---
+category: Fixed
+---
+
+- **Chat card update evidence** — distinguishes an accepted update request from an independently verified visible update, preserving the real `bizId` and warning callers not to repeat an unverified write.
+- **Chat command guidance** — splits message and group references by task and explains that `--from` is ambiguous between sender and time-range intent.

.changes/released/1.0.59-beta.2/schema-catalog-json-projection.md

@@ -0,0 +1,8 @@
+---
+category: Changed
+---
+
+- **Faster Schema Catalog assembly** — projects typed values into payload JSON
+  without re-running a validation scan over documents `json.Marshal` has just
+  produced, cutting roughly a third of the projection work across the full tool
+  set. Untrusted JSON input keeps its existing validation.

.changes/released/1.0.59-beta.2/wiki-shortcut-alignment.md

@@ -0,0 +1,9 @@
+---
+category: Added
+---
+
+- **Wiki Shortcut workflows** — publishes 20 reviewed space, member, node, and
+  activity shortcuts with strict collection validation, cursor handling,
+  write-terminal evidence, safe read-backs where the backend supports them,
+  task-oriented routing, and documented backend
+  boundaries.

.changes/released/1.0.59-beta.3/1006-aitable-pagination-minutes-unshare.md

@@ -0,0 +1,11 @@
+---
+category: Fixed
+---
+
+- **Aitable pagination and Minutes unshare verification** (#1006) — keeps
+  record queries on the service's 20-record page boundary so multi-page reads
+  and mutation readbacks no longer report false retryable failures, preserves
+  `totalCount` when supplied, validates `--dry-run` plans before transport,
+  follows active deletion readback continuations before proving absence, and
+  rejects Minutes unshare success until the listening note exists and the
+  service acknowledges the exact task and member targets.

.changes/released/1.0.59-beta.3/928-chat-bot-group-reply.md

@@ -0,0 +1,5 @@
+---
+category: Added
+---
+
+- **Robot group reference replies** (#928) — `chat message send-by-bot` supports paired `--reply` and `--ref-sender` flags for Markdown replies that quote an existing group message.

.changes/released/1.0.59-beta.3/960-doc-readback-verification.md

@@ -0,0 +1,5 @@
+---
+category: Fixed
+---
+
+- **Document write verification** (#960) — avoids false partial-success results when normalized Markdown, paginated blocks, inline images, or version reverts are confirmed by server readback. Document reverts and media inserts now require explicit readback evidence and report partial success when the server cannot prove the requested result.

.changes/released/1.0.59-beta.3/aitable-param-aliases.md

@@ -0,0 +1,5 @@
+---
+category: Changed
+---
+
+- **AI Table parameter aliases** — accepts reviewed equivalent spellings for Base, table, workflow, search, pagination, and description parameters while keeping role-changing or semantically ambiguous inputs blocked.

.changes/released/1.0.59-beta.3/aitable-record-stats.md

@@ -0,0 +1,9 @@
+---
+category: Added
+---
+
+- **AI Table server-side statistics** — adds `dws aitable record stats` for
+  ungrouped record-set metrics through `query_records_stats`, plus `dws aitable
+  record group-stats` for grouped, distinct, and advanced aggregation through
+  `query_stats`; both commands validate their JSON aggregation contracts before
+  dispatch.

.changes/released/1.0.59-beta.3/calendar-event-share-info.md

@@ -0,0 +1,5 @@
+---
+category: Added
+---
+
+- **Calendar event share-info** (#980) — adds `dws calendar event share-info` to fetch a calendar event's share info (title, organizer, location, join info) for sharing with others; supports `--calendar-id` and `--language`.

.changes/released/1.0.59-beta.3/calendar-todo-shortcut-alignment.md

@@ -0,0 +1,11 @@
+---
+category: Added
+---
+
+- **Calendar and To-do Shortcut workflows** — aligns 47 public task-oriented
+  entries with lark-cli where the DingTalk backend supports equivalent
+  semantics, rejects malformed or missing collections instead of returning
+  false empty success, preserves truthful pagination, and requires stable
+  identifiers plus read-back or explicit terminal receipts for writes. Adds
+  deterministic contract coverage, a PII-safe live E2E runner, and a sanitized
+  capability review with documented platform boundaries.

.changes/released/1.0.59-beta.3/chat-sender-identity-contract.md

@@ -0,0 +1,5 @@
+---
+category: Fixed
+---
+
+- **Chat sender identity guards** — preserves unverified mixed sender inputs after exact message `senderId` matches and aligns `--sender-query` Skill guidance with fail-closed Runtime behavior.

.changes/released/1.0.59-beta.3/doc-drive-description-scope.md

@@ -0,0 +1,5 @@
+---
+category: Changed
+---
+
+- **Doc/drive description scope** — restates the `dingtalk-doc` description as document-entity-and-content operations with an explicit exclusion list, and narrows `dingtalk-drive` to file-level management of DingTalk documents, so first-round Agent selection separates content work from file management without changing CLI behavior.

.changes/released/1.0.59-beta.3/doc-sheet-comment-lifecycle.md

@@ -0,0 +1,10 @@
+---
+category: Added
+---
+
+- **Doc and Sheet comment lifecycle commands** — adds `comment batch-query`,
+  `comment resolve`, `comment restore`, and the lightweight
+  `comment react-reply` to both `dws doc` and `dws sheet`. The two domains share
+  the same `doc-comment` MCP capabilities; batch queries preserve input order
+  for repeated `topicId:commentKey` references, while reaction replies require
+  DingTalk reaction names such as `憨笑` or `鼓掌` rather than raw Unicode emoji.

.changes/released/1.0.59-beta.3/sheet-source-range-dropdowns.md

@@ -0,0 +1,6 @@
+---
+category: Added
+---
+
+- **Sheet SourceRange dropdowns** — supports range-backed dropdowns across direct, cell, and batch write paths, with structured readback for valid and invalid references. Batch `set-dropdown` now rejects unsupported top-level `colors` / `source-colors`; Inline colors belong in `options[].color`, while SourceRange color writes remain unsupported.
+- **Sheet read completion metadata** — documents and preserves returned ranges, truncation reasons, and partial-read status for large range and CSV reads.

.changes/released/1.0.59-beta.3/windows-event-bus-lifecycle.md

@@ -0,0 +1,5 @@
+---
+category: Fixed
+---
+
+- **Windows event bus lifecycle** — start event consumers without unsupported inherited file descriptors, stop buses through local IPC with a termination fallback, and preserve subscription cleanup when startup fails.

.changes/wait-framework-capability.md

@@ -0,0 +1,26 @@
+---
+category: Added
+---
+
+- **Wait framework capability** — adds the reviewed `Contract.Wait`
+  declaration (`contract.WaitSpec`) with three execution modes: `poll`
+  (cadence-poll the leaf's `WaitPoll` hook), `event` (consume the leaf's
+  `WaitEvents` push stream, correlate events to the accepted resource via
+  `match_field`/`resource_query`, apply the same terminal map), and `auto`
+  (event first, fall back to polling when the stream ends or the
+  subscription fails — one deadline spans both phases). Declared commands
+  must use the `ResultInvoke` dispatcher; mode and hooks are paired at
+  construction (poll↔WaitPoll, event↔WaitEvents, auto↔both; surplus hooks
+  are rejected too). Declared commands register `--wait` /
+  `--wait-timeout` (framework-owned flags that never enter MCP toolArgs);
+  undeclared commands reject the flags as unknown. The wait phase closes
+  the unified envelope exactly once: terminal success → `success`,
+  terminal failure → `failure` with new wire-stable `error.type: "wait"`
+  (exit code 8), timeout → `pending` with `meta.operation.timed_out: true`
+  and the last observed state (exit 0). Deadline exhaustion during a poll,
+  during event consumption, or between polls always closes as timed-out
+  pending, never as a poll/stream failure; a correlated event with an
+  unknown status fails closed exactly like a poll. The capability is
+  projected into the Schema catalog (`wait` key) alongside `dry_run`. No
+  business command declares it yet; approval/export/batch adoption lands
+  separately.

.github/workflows/ci.yml

@@ -151,6 +151,10 @@ jobs:
                 filename.startsWith('scripts/') ||
                 filename.startsWith('verify/') ||
                 filename.startsWith('internal/helpers/') ||
+                // Shortcut declarations feed the live command tree and Schema
+                // assembly. Their reverse dependencies include the expensive
+                // app and generator packages, which must run in separate shards.
+                filename.startsWith('internal/shortcut/') ||
                 filename.startsWith('internal/generator/') ||
                 filename.startsWith('internal/cli/schema') ||
                 // Parameter aliases are reduced against the live command tree.
@@ -509,11 +513,38 @@ jobs:
         run: node .github/reviewer-routing.test.js
 
   test-focused:
-    name: Test (changed packages)
+    name: "Test (focused: ${{ matrix.shard }})"
     needs: lint
     if: ${{ needs.lint.outputs.changelog_only != 'true' && needs.lint.outputs.docs_only != 'true' && needs.lint.outputs.full_suite != 'true' }}
     runs-on: ubuntu-latest
+    # Each shard owns one bounded slice of the impacted set, so no single job
+    # carries internal/app together with every reverse dependency. The shard
+    # list and per-shard execution below mirror test-race, which runs the same
+    # shards at full-suite scope; release-scripts is included because its
+    # dedicated job only runs at full-suite or release-sensitive scope, and
+    # dropping it here would stop testing test/scripts changes entirely.
+    # internal/app is carried by one shard per bounded partition rather than a
+    # single app shard: the partitions used to run end to end inside one job,
+    # where the Schema partition alone owned most of the wall clock. The
+    # app-<partition> names are pinned to the helper's partition set by
+    # TestCIAppRacePartitionMatrixMatchesHelper, so a partition can never lose
+    # its job silently.
     timeout-minutes: 20
+    strategy:
+      fail-fast: false
+      matrix:
+        shard:
+          - app-schema
+          - app-a-b
+          - app-c
+          - app-d-r
+          - app-s-z-example-fuzz
+          - generators
+          - helpers
+          - cli
+          - smoke
+          - remaining
+          - release-scripts
     steps:
       - name: Check out repository
         uses: actions/checkout@v4
@@ -542,37 +573,110 @@ jobs:
         with:
           go-version-file: go.mod
 
-      - name: Test changed packages and reverse dependencies
+      - name: Select impacted packages for shard
+        id: select
+        shell: bash
+        env:
+          TEST_SHARD: ${{ matrix.shard }}
+        run: |
+          set -euo pipefail
+          # Every app partition shard tests the same single internal/app
+          # package, so the impacted-package query uses the base shard name and
+          # the partition only selects which tests run.
+          package_shard="$TEST_SHARD"
+          case "$TEST_SHARD" in
+            app-*) package_shard=app ;;
+          esac
+          package_output="$(
+            ./scripts/ci/changed-test-packages.sh \
+              list-shard "$package_shard" "$TEST_BASE_REF" "$TEST_HEAD_REF"
+          )"
+          if [ -z "$package_output" ]; then
+            echo "No buildable Go package in shard $TEST_SHARD is affected by this revision." \
+              >> "$GITHUB_STEP_SUMMARY"
+            echo "affected=false" >> "$GITHUB_OUTPUT"
+            exit 0
+          fi
+          # The package list travels through a file rather than a step output:
+          # reading it with `mapfile < file` has unambiguous line semantics,
+          # whereas a here-string over a multi-line output would append an extra
+          # empty element if the value ever carried a trailing newline, and an
+          # empty element would reach go test as an empty package argument.
+          printf '%s\n' "$package_output" > "$RUNNER_TEMP/focused-shard-packages.txt"
+          echo "affected=true" >> "$GITHUB_OUTPUT"
+
+      - name: Build
+        if: ${{ matrix.shard == 'remaining' && steps.select.outputs.affected == 'true' }}
+        run: make build
+
+      - name: Install archive tooling
+        if: ${{ matrix.shard == 'release-scripts' && steps.select.outputs.affected == 'true' }}
+        run: sudo apt-get update && sudo apt-get install -y zip unzip
+
+      - name: Test shard with Race Detection
+        if: ${{ steps.select.outputs.affected == 'true' }}
         shell: bash
         env:
           DWS_PACKAGE_VERSION: 0.0.0-test
+          TEST_SHARD: ${{ matrix.shard }}
         run: |
           set -euo pipefail
-          package_output="$(
-            ./scripts/ci/changed-test-packages.sh \
-              list "$TEST_BASE_REF" "$TEST_HEAD_REF"
-          )"
-          if [ -z "$package_output" ]; then
-            echo "No buildable Go package is affected by this revision." \
-              >> "$GITHUB_STEP_SUMMARY"
+          mapfile -t packages < "$RUNNER_TEMP/focused-shard-packages.txt"
+          test "${#packages[@]}" -gt 0
+          for package in "${packages[@]}"; do
+            test -n "$package" || {
+              echo "shard package list contains an empty entry" >&2
+              exit 1
+            }
+          done
+          case "$TEST_SHARD" in
+            app-*)
+              # A single long-lived app test process retains every constructed
+              # command tree in framework registries. Each partition is its own
+              # job, so that state is released when the process exits and the
+              # partitions run concurrently instead of end to end. The helper
+              # still verifies that the partition patterns cover every top-level
+              # test exactly once before running the one it was asked for.
+              test "${#packages[@]}" -eq 1
+              ./scripts/ci/run-app-race-tests.sh run "${packages[0]}" "${TEST_SHARD#app-}"
+              exit 0
+              ;;
+          esac
+          if [ "$TEST_SHARD" = "release-scripts" ]; then
+            # Mirror the dedicated release-contract job: these suites shell out
+            # to archive tooling and are not race-instrumented there.
+            go test -v -count=1 -timeout=10m "${packages[@]}"
             exit 0
           fi
-          mapfile -t packages <<< "$package_output"
-          go test -v -race -count=1 -timeout=15m "${packages[@]}"
+          # cli/smoke own heavy NewRootCommand / Schema assembly under -race;
+          # give them a dedicated package timeout on slower hosted runners.
+          timeout_budget=12m
+          if [ "$TEST_SHARD" = "cli" ] ||
+             [ "$TEST_SHARD" = "smoke" ]; then
+            timeout_budget=15m
+          fi
+          go test -v -race -count=1 -timeout="$timeout_budget" "${packages[@]}"
 
   test-race:
     name: "Test (race: ${{ matrix.shard }})"
     needs: lint
     if: ${{ needs.lint.outputs.changelog_only != 'true' && needs.lint.outputs.docs_only != 'true' && needs.lint.outputs.full_suite == 'true' }}
     runs-on: ubuntu-latest
-    # app runs several independently bounded processes; cli/smoke need headroom
-    # beyond go test -timeout for setup + assembly.
+    # internal/app is split across one shard per bounded partition so the
+    # partitions run concurrently and each releases its framework registries
+    # when the process exits; cli/smoke need headroom beyond go test -timeout for
+    # setup + assembly. The app-<partition> names are pinned to the helper's
+    # partition set by TestCIAppRacePartitionMatrixMatchesHelper.
     timeout-minutes: 20
     strategy:
       fail-fast: false
       matrix:
         shard:
-          - app
+          - app-schema
+          - app-a-b
+          - app-c
+          - app-d-r
+          - app-s-z-example-fuzz
           - generators
           - helpers
           - cli
@@ -598,18 +702,30 @@ jobs:
           TEST_SHARD: ${{ matrix.shard }}
         run: |
           set -euo pipefail
-          package_output="$(./scripts/ci/test-packages.sh list "$TEST_SHARD")"
+          # Every app partition shard tests the same single internal/app
+          # package, so the package query uses the base shard name and the
+          # partition only selects which tests run.
+          package_shard="$TEST_SHARD"
+          case "$TEST_SHARD" in
+            app-*) package_shard=app ;;
+          esac
+          package_output="$(./scripts/ci/test-packages.sh list "$package_shard")"
           test -n "$package_output"
           mapfile -t packages <<< "$package_output"
           test "${#packages[@]}" -gt 0
-          if [ "$TEST_SHARD" = "app" ]; then
-            # A single long-lived app test process retains every constructed
-            # command tree in framework registries. Isolate Schema assembly and
-            # bounded name ranges so each process releases that state on exit.
-            test "${#packages[@]}" -eq 1
-            ./scripts/ci/run-app-race-tests.sh run "${packages[0]}"
-            exit 0
-          fi
+          case "$TEST_SHARD" in
+            app-*)
+              # A single long-lived app test process retains every constructed
+              # command tree in framework registries. Each partition is its own
+              # job, so that state is released when the process exits and the
+              # partitions run concurrently instead of end to end. The helper
+              # still verifies that the partition patterns cover every top-level
+              # test exactly once before running the one it was asked for.
+              test "${#packages[@]}" -eq 1
+              ./scripts/ci/run-app-race-tests.sh run "${packages[0]}" "${TEST_SHARD#app-}"
+              exit 0
+              ;;
+          esac
           # cli/smoke own heavy NewRootCommand / Schema assembly under -race;
           # give them a dedicated package timeout on slower hosted runners.
           timeout_budget=12m
@@ -711,7 +827,7 @@ jobs:
           failed=0
           if [ "$CHANGELOG_ONLY" = true ] || [ "$DOCS_ONLY" = true ]; then
             for shard in \
-              "changed packages:$FOCUSED_RESULT" \
+              "focused shards:$FOCUSED_RESULT" \
               "race shards:$RACE_RESULT" \
               "release scripts:$RELEASE_SCRIPTS_RESULT" \
               "cross-platform compile:$CROSS_PLATFORM_RESULT" \
@@ -740,7 +856,7 @@ jobs:
             release_expected=success
           fi
           for shard in \
-            "changed packages:$FOCUSED_RESULT:$focused_expected" \
+            "focused shards:$FOCUSED_RESULT:$focused_expected" \
             "race shards:$RACE_RESULT:$race_expected" \
             "release scripts:$RELEASE_SCRIPTS_RESULT:$release_expected" \
             "cross-platform compile:$CROSS_PLATFORM_RESULT:success"

.gitignore

@@ -20,6 +20,10 @@ test/cli_compat/testdata/
 .gitignore
 .worktrees/
 .qoder/
+_logs/
+_docs/
+_output/
+vendor/
 
 # Secrets & credentials
 .env

CHANGELOG.md

@@ -6,6 +6,92 @@ The format is inspired by [Keep a Changelog](https://keepachangelog.com/) and th
 
 ## [Unreleased]
 
+## [1.0.59-beta.3] - 2026-08-19
+
+### Added
+
+- **Robot group reference replies** (#928) — `chat message send-by-bot` supports paired `--reply` and `--ref-sender` flags for Markdown replies that quote an existing group message.
+
+- **AI Table server-side statistics** — adds `dws aitable record stats` for
+  ungrouped record-set metrics through `query_records_stats`, plus `dws aitable
+  record group-stats` for grouped, distinct, and advanced aggregation through
+  `query_stats`; both commands validate their JSON aggregation contracts before
+  dispatch.
+
+- **Calendar event share-info** (#980) — adds `dws calendar event share-info` to fetch a calendar event's share info (title, organizer, location, join info) for sharing with others; supports `--calendar-id` and `--language`.
+
+- **Calendar and To-do Shortcut workflows** — aligns 47 public task-oriented
+  entries with lark-cli where the DingTalk backend supports equivalent
+  semantics, rejects malformed or missing collections instead of returning
+  false empty success, preserves truthful pagination, and requires stable
+  identifiers plus read-back or explicit terminal receipts for writes. Adds
+  deterministic contract coverage, a PII-safe live E2E runner, and a sanitized
+  capability review with documented platform boundaries.
+
+- **Doc and Sheet comment lifecycle commands** — adds `comment batch-query`,
+  `comment resolve`, `comment restore`, and the lightweight
+  `comment react-reply` to both `dws doc` and `dws sheet`. The two domains share
+  the same `doc-comment` MCP capabilities; batch queries preserve input order
+  for repeated `topicId:commentKey` references, while reaction replies require
+  DingTalk reaction names such as `憨笑` or `鼓掌` rather than raw Unicode emoji.
+
+- **Sheet SourceRange dropdowns** — supports range-backed dropdowns across direct, cell, and batch write paths, with structured readback for valid and invalid references. Batch `set-dropdown` now rejects unsupported top-level `colors` / `source-colors`; Inline colors belong in `options[].color`, while SourceRange color writes remain unsupported.
+- **Sheet read completion metadata** — documents and preserves returned ranges, truncation reasons, and partial-read status for large range and CSV reads.
+
+### Changed
+
+- **AI Table parameter aliases** — accepts reviewed equivalent spellings for Base, table, workflow, search, pagination, and description parameters while keeping role-changing or semantically ambiguous inputs blocked.
+
+- **Doc/drive description scope** — restates the `dingtalk-doc` description as document-entity-and-content operations with an explicit exclusion list, and narrows `dingtalk-drive` to file-level management of DingTalk documents, so first-round Agent selection separates content work from file management without changing CLI behavior.
+
+### Fixed
+
+- **Aitable pagination and Minutes unshare verification** (#1006) — keeps
+  record queries on the service's 20-record page boundary so multi-page reads
+  and mutation readbacks no longer report false retryable failures, preserves
+  `totalCount` when supplied, validates `--dry-run` plans before transport,
+  follows active deletion readback continuations before proving absence, and
+  rejects Minutes unshare success until the listening note exists and the
+  service acknowledges the exact task and member targets.
+
+- **Document write verification** (#960) — avoids false partial-success results when normalized Markdown, paginated blocks, inline images, or version reverts are confirmed by server readback. Document reverts and media inserts now require explicit readback evidence and report partial success when the server cannot prove the requested result.
+
+- **Chat sender identity guards** — preserves unverified mixed sender inputs after exact message `senderId` matches and aligns `--sender-query` Skill guidance with fail-closed Runtime behavior.
+
+- **Windows event bus lifecycle** — start event consumers without unsupported inherited file descriptors, stop buses through local IPC with a termination fallback, and preserve subscription cleanup when startup fails.
+
+
+## [1.0.59-beta.2] - 2026-08-17
+
+### Added
+
+- **Privacy-safe CLI telemetry** (#1009) — reports reviewed command outcomes and profile identity dimensions while excluding command arguments, output, paths, device fingerprints, and automatic system dimensions; `DO_NOT_TRACK=1` disables reporting.
+
+- **Feedback survey entry in root help** (#1019) — `dws --help` now closes with a Feedback section linking the user-experience survey form.
+
+- **Wiki Shortcut workflows** — publishes 20 reviewed space, member, node, and
+  activity shortcuts with strict collection validation, cursor handling,
+  write-terminal evidence, safe read-backs where the backend supports them,
+  task-oriented routing, and documented backend
+  boundaries.
+
+### Changed
+
+- **Chat IM ID flags** (#954) — standardizes chat command entry points on `--conversation-id` for conversation IDs and `--message-id` for message IDs, so help, Schema, and Agent recommendations use the same canonical flags.
+- **Legacy chat flag compatibility** (#954) — keeps older chat IM ID flags such as `--group`, `--id`, `--chat`, `--open-conversation-id`, `--msg-id`, and `--open-message-id` working as compatibility aliases where applicable, while hiding migrated aliases from recommended help and Schema surfaces.
+- **Chat group bots target flag** (#954) — keeps `dws chat group bots` on the visible `--group` flag; this command does not register `--group-name`, and `--group` accepts either an openConversationId or a uniquely resolved group name.
+
+- **Faster Schema Catalog assembly** — projects typed values into payload JSON
+  without re-running a validation scan over documents `json.Marshal` has just
+  produced, cutting roughly a third of the projection work across the full tool
+  set. Untrusted JSON input keeps its existing validation.
+
+### Fixed
+
+- **Chat card update evidence** — distinguishes an accepted update request from an independently verified visible update, preserving the real `bizId` and warning callers not to repeat an unverified write.
+- **Chat command guidance** — splits message and group references by task and explains that `--from` is ambiguous between sender and time-range intent.
+
+
 ## [1.0.59-beta.1] - 2026-08-14
 
 ### Added

Formula/dingtalk-workspace-cli-beta.rb

@@ -1,33 +1,33 @@
 class DingtalkWorkspaceCliBeta < Formula
   desc "Automate DingTalk workspace tasks from the terminal (beta channel)"
   homepage "https://github.com/DingTalk-Real-AI/dingtalk-workspace-cli"
-  version "1.0.59-beta.1"
+  version "1.0.59-beta.3"
   license "Apache-2.0"
   keg_only "it is the beta channel and conflicts with dingtalk-workspace-cli"
 
   on_macos do
     if Hardware::CPU.arm?
-      url "https://github.com/DingTalk-Real-AI/dingtalk-workspace-cli/releases/download/v1.0.59-beta.1/dws-darwin-arm64.tar.gz"
-      sha256 "36a30f3496e0f759c15c0b09f67dbd23b8ecdfff2eebe572f88125b26485830f"
+      url "https://github.com/DingTalk-Real-AI/dingtalk-workspace-cli/releases/download/v1.0.59-beta.3/dws-darwin-arm64.tar.gz"
+      sha256 "9c99adcefd9104368eb443f0a1b4af8e7aceaa1ffdd4462e486854c1692bb6ce"
     else
-      url "https://github.com/DingTalk-Real-AI/dingtalk-workspace-cli/releases/download/v1.0.59-beta.1/dws-darwin-amd64.tar.gz"
-      sha256 "e7a04906380efd8da88cd112e6a512bb6470a3956dc370150037ed6e314db445"
+      url "https://github.com/DingTalk-Real-AI/dingtalk-workspace-cli/releases/download/v1.0.59-beta.3/dws-darwin-amd64.tar.gz"
+      sha256 "f5cc8efb1f982d68ae549190fd683292359c2ab542b532fa52bb35e6b5c049af"
     end
   end
 
   on_linux do
     if Hardware::CPU.arm?
-      url "https://github.com/DingTalk-Real-AI/dingtalk-workspace-cli/releases/download/v1.0.59-beta.1/dws-linux-arm64.tar.gz"
-      sha256 "f59ab055f3e841e4cebc964ae3ef969668475548abaaf8bede44afdca9a3e28d"
+      url "https://github.com/DingTalk-Real-AI/dingtalk-workspace-cli/releases/download/v1.0.59-beta.3/dws-linux-arm64.tar.gz"
+      sha256 "7a4efd04b417ce8013b1e431274b396179958da244164f59974358ba327ff093"
     else
-      url "https://github.com/DingTalk-Real-AI/dingtalk-workspace-cli/releases/download/v1.0.59-beta.1/dws-linux-amd64.tar.gz"
-      sha256 "2c8f919489d958c7d49262615e81faac70a9fbcae2d589ab54a0bb3c5700a057"
+      url "https://github.com/DingTalk-Real-AI/dingtalk-workspace-cli/releases/download/v1.0.59-beta.3/dws-linux-amd64.tar.gz"
+      sha256 "90181e8f2e9010c1943a5773c3d45d7d3ac85d6bc93e18a9aaa7c69909e553d7"
     end
   end
 
   resource "skills" do
-    url "https://github.com/DingTalk-Real-AI/dingtalk-workspace-cli/releases/download/v1.0.59-beta.1/dws-skills.zip"
-    sha256 "25f4a7e1d01fa4d771d79201b34b11ee8a24182bdcdc94bfb98d2bd5845bed3b"
+    url "https://github.com/DingTalk-Real-AI/dingtalk-workspace-cli/releases/download/v1.0.59-beta.3/dws-skills.zip"
+    sha256 "e7028914a4a826af9b18ed4922d68fa8f279473817fed4f305465bc8a7aad363"
   end
 
   def install

README.md

@@ -482,7 +482,7 @@ Env vars: `DWS_SKILL_MODE=mono|multi` (also honored by `install.sh` / `install.p
 <details>
 <summary><strong>Personal Event Subscription</strong> — real-time DingTalk messages for event-driven agents</summary>
 
-`dws event consume` subscribes as the currently logged-in user over a managed Stream WebSocket and emits each event as one NDJSON line on stdout. The public catalog covers scoped and all one-to-one/group messages, specified senders, read/recall/reaction events, group lifecycle events, and six OA approval task/instance events.
+`dws event consume` subscribes as the currently logged-in user over a managed Stream WebSocket and emits each event as one NDJSON line on stdout. The public catalog covers scoped and all one-to-one/group messages, specified senders, read/recall/reaction events, group lifecycle events, and seven OA approval task/instance events.
 
 The default `ndjson`, `json`, and `pretty` output preserves the transport envelope (`type`, `event_type`, string `data`, and `headers`) for existing scripts; `compact` retains its existing processor. Add `--flatten` to emit the stable top-level business fields used by Agent workflows. `--format` controls JSON serialization; `--flatten` controls the data structure and cannot be combined with `-f raw` or `--debug-raw-events`.
 
@@ -530,12 +530,13 @@ dws event consume user_im_group_disbanded --group <openConversationId> --flatten
 dws event +listen-im --kind sender --user <userId> \
   --events message,read,recall -f ndjson
 
-# Listen for all six public OA approval events in one process
+# Listen for all seven public OA approval events in one process
 dws event consume \
   user_oa_approval_task_created \
   user_oa_approval_task_finished \
   user_oa_approval_task_redirected \
   user_oa_approval_instance_started \
+  user_oa_approval_instance_cc \
   user_oa_approval_instance_terminated \
   user_oa_approval_instance_finished \
   --flatten -f ndjson

README_zh.md

@@ -476,7 +476,7 @@ multi setup 或 upgrade 后，DWS 会把官方 bundle 快照和统一所有权
 <details>
 <summary><strong>个人事件订阅</strong> — 实时接收钉钉消息，驱动事件触发的 Agent</summary>
 
-`dws event consume` 使用当前 OAuth 登录用户建立托管的 Stream WebSocket 长连接，并把每条事件以 NDJSON 一行输出到 stdout。当前公开目录覆盖指定范围和全量单聊/群消息、指定发送人、已读/撤回/表情回应、群生命周期，以及六个 OA 审批任务/实例事件。
+`dws event consume` 使用当前 OAuth 登录用户建立托管的 Stream WebSocket 长连接，并把每条事件以 NDJSON 一行输出到 stdout。当前公开目录覆盖指定范围和全量单聊/群消息、指定发送人、已读/撤回/表情回应、群生命周期，以及七个 OA 审批任务/实例事件。
 
 默认 `ndjson`、`json`、`pretty` 输出保留兼容 transport envelope（`type`、`event_type`、字符串 `data`、`headers`），`compact` 继续沿用原 processor。Agent 或新脚本显式加 `--flatten` 后，输出稳定的顶层业务字段。`--format` 控制 JSON 序列化，`--flatten` 控制数据结构，且不能与 `-f raw` 或 `--debug-raw-events` 同时使用。
 
@@ -524,12 +524,13 @@ dws event consume user_im_group_disbanded --group <openConversationId> --flatten
 dws event +listen-im --kind sender --user <userId> \
   --events message,read,recall -f ndjson
 
-# 一个进程监听全部六个公开 OA 审批事件
+# 一个进程监听全部七个公开 OA 审批事件
 dws event consume \
   user_oa_approval_task_created \
   user_oa_approval_task_finished \
   user_oa_approval_task_redirected \
   user_oa_approval_instance_started \
+  user_oa_approval_instance_cc \
   user_oa_approval_instance_terminated \
   user_oa_approval_instance_finished \
   --flatten -f ndjson

cmd/main.go

@@ -15,12 +15,76 @@ package main
 
 import (
 	"os"
+	"strings"
 
 	"github.com/DingTalk-Real-AI/dingtalk-workspace-cli/internal/app"
+	"gitlab.alibaba-inc.com/aes/aem-go-sdk/clitrack"
 )
 
-var exit = os.Exit
+var (
+	appExecute               = app.ExecuteWithTelemetry
+	resolveTelemetryIdentity = app.ResolveTelemetryIdentity
+	trackRun                 = func(cfg clitrack.Config, execute func() error, exitCode func(error) int) {
+		clitrack.New(cfg).Run(execute, exitCode)
+	}
+)
+
+// trackedExitError tells clitrack that the command failed without asking it to
+// print the error a second time. The already-rendered message is published via
+// ExtraFields c5, while app.Execute remains the sole owner of presentation.
+type trackedExitError struct{}
+
+func (trackedExitError) Error() string { return "" }
+
+func trackerConfig(identity app.TelemetryIdentity, commandPath, errorMessage *string) clitrack.Config {
+	return clitrack.Config{
+		PID:                   "wcCRwZ",
+		App:                   "dws",
+		Version:               app.RawVersion(),
+		UID:                   identity.UserID,
+		Username:              identity.UserName,
+		NoCommandLine:         true,
+		NoCwd:                 true,
+		NoAutomaticDimensions: true,
+		ExtraFields: func() map[string]string {
+			fields := map[string]string{"c9": *commandPath}
+			if identity.CorpID != "" {
+				fields["c10"] = identity.CorpID
+			}
+			if *errorMessage != "" {
+				fields["c5"] = *errorMessage
+			}
+			return fields
+		},
+	}
+}
+
+func telemetryOptedOut() bool {
+	return strings.TrimSpace(os.Getenv("DO_NOT_TRACK")) != ""
+}
 
 func main() {
-	exit(app.Execute())
+	optedOut := telemetryOptedOut()
+	identity := app.TelemetryIdentity{}
+	if !optedOut {
+		identity = resolveTelemetryIdentity(os.Args[1:])
+	}
+	exitCode := 0
+	commandPath := "dws"
+	errorMessage := ""
+	cfg := trackerConfig(identity, &commandPath, &errorMessage)
+	if optedOut {
+		cfg.PID = ""
+	}
+	trackRun(
+		cfg,
+		func() error {
+			exitCode, commandPath, errorMessage = appExecute()
+			if exitCode != 0 {
+				return trackedExitError{}
+			}
+			return nil
+		},
+		func(error) int { return exitCode },
+	)
 }

cmd/main_test.go

@@ -1,27 +1,220 @@
 package main
 
 import (
+	"encoding/json"
+	"fmt"
+	"io"
+	"net/http"
+	"net/http/httptest"
+	"net/url"
 	"os"
+	"slices"
+	"sort"
+	"strings"
 	"testing"
+	"time"
+
+	"github.com/DingTalk-Real-AI/dingtalk-workspace-cli/internal/app"
+	"github.com/DingTalk-Real-AI/dingtalk-workspace-cli/internal/testseam"
+	"gitlab.alibaba-inc.com/aes/aem-go-sdk/clitrack"
 )
 
-func TestCrossPlatformCoverageMainExitsWithSuccessfulVersionCommand(t *testing.T) {
-	previousExit := exit
-	previousArgs := os.Args
-	t.Cleanup(func() {
-		exit = previousExit
-		os.Args = previousArgs
-	})
+func TestCrossPlatformCoverageMainRunsThroughCLITracker(t *testing.T) {
+	for _, wantCode := range []int{0, 1, 3, 5} {
+		t.Run(fmt.Sprintf("exit_%d", wantCode), func(t *testing.T) {
+			t.Setenv("DO_NOT_TRACK", "")
+			wantError := ""
+			if wantCode != 0 {
+				wantError = "synthetic failure"
+			}
+			testseam.Swap(t, &os.Args, []string{"dws", "sheet", "read", "--profile", "corp-a"})
+			testseam.Swap(t, &resolveTelemetryIdentity, func(args []string) app.TelemetryIdentity {
+				if strings.Join(args, " ") != "sheet read --profile corp-a" {
+					t.Fatalf("telemetry identity args = %#v", args)
+				}
+				return app.TelemetryIdentity{UserID: "user-1", UserName: "Alice", CorpID: "corp-1"}
+			})
+			testseam.Swap(t, &appExecute, func() (int, string, string) { return wantCode, "sheet read", wantError })
+			called := false
+			testseam.Swap(t, &trackRun, func(cfg clitrack.Config, execute func() error, exitCode func(error) int) {
+				called = true
+				if cfg.PID != "wcCRwZ" || cfg.App != "dws" {
+					t.Fatalf("tracker identity = PID %q App %q", cfg.PID, cfg.App)
+				}
+				if cfg.Version != app.RawVersion() {
+					t.Fatalf("tracker Version = %q, want %q", cfg.Version, app.RawVersion())
+				}
+				if !cfg.NoCommandLine || !cfg.NoCwd || !cfg.NoAutomaticDimensions || cfg.CaptureOutput {
+					t.Fatalf("tracker privacy config = NoCommandLine %v NoCwd %v NoAutomaticDimensions %v CaptureOutput %v", cfg.NoCommandLine, cfg.NoCwd, cfg.NoAutomaticDimensions, cfg.CaptureOutput)
+				}
+				if cfg.Env != "" || cfg.EventID != "" || cfg.Endpoint != "" || cfg.FlushTimeout != 0 || cfg.OutputMaxLen != 0 {
+					t.Fatalf("tracker SDK defaults were overridden: %#v", cfg)
+				}
+				if cfg.UID != "user-1" || cfg.Username != "Alice" || cfg.UserType != "" {
+					t.Fatalf("tracker user identity = UID %q Username %q UserType %q", cfg.UID, cfg.Username, cfg.UserType)
+				}
 
+				err := execute()
+				if wantCode == 0 && err != nil {
+					t.Fatalf("successful tracked execute error = %v", err)
+				}
+				if wantCode != 0 && (err == nil || err.Error() != "") {
+					t.Fatalf("failed tracked execute error = %#v, want empty sentinel", err)
+				}
+				if gotCode := exitCode(err); gotCode != wantCode {
+					t.Fatalf("tracked exit code = %d, want %d", gotCode, wantCode)
+				}
+				fields := cfg.ExtraFields()
+				if fields["c9"] != "sheet read" || fields["c10"] != "corp-1" || fields["c5"] != wantError {
+					t.Fatalf("tracker extra fields = %#v, want command path, corp ID, and error %q", fields, wantError)
+				}
+				if (wantError == "" && len(fields) != 2) || (wantError != "" && len(fields) != 3) {
+					t.Fatalf("tracker extra field count = %d for error %q", len(fields), wantError)
+				}
+			})
+
+			main()
+			if !called {
+				t.Fatalf("trackRun was not called for exit code %d", wantCode)
+			}
+		})
+	}
+}
+
+func TestCrossPlatformCoverageTrackerConfigOmitsEmptyOrganization(t *testing.T) {
+	commandPath := "version"
+	errorMessage := ""
+	cfg := trackerConfig(app.TelemetryIdentity{}, &commandPath, &errorMessage)
+	if cfg.UID != "" {
+		t.Fatalf("empty identity UID = %q", cfg.UID)
+	}
+	if cfg.Username != "" {
+		t.Fatalf("empty identity Username = %q", cfg.Username)
+	}
+	if fields := cfg.ExtraFields(); len(fields) != 1 || fields["c9"] != "version" {
+		t.Fatalf("empty organization fields = %#v", fields)
+	}
+}
+
+func TestCrossPlatformCoverageDefaultTrackRunNoopTracker(t *testing.T) {
 	called := false
-	code := -1
-	exit = func(value int) {
+	trackRun(clitrack.Config{}, func() error {
 		called = true
-		code = value
+		return nil
+	}, nil)
+	if !called {
+		t.Fatal("default tracker did not execute callback")
 	}
-	os.Args = []string{"dws", "version"}
+}
+
+func TestCrossPlatformCoverageMainRespectsDoNotTrack(t *testing.T) {
+	t.Setenv("DO_NOT_TRACK", "1")
+	testseam.Swap(t, &os.Args, []string{"dws", "version"})
+	testseam.Swap(t, &resolveTelemetryIdentity, func([]string) app.TelemetryIdentity {
+		t.Fatal("DO_NOT_TRACK must skip telemetry identity reads")
+		return app.TelemetryIdentity{}
+	})
+	testseam.Swap(t, &appExecute, func() (int, string, string) { return 0, "version", "" })
+	testseam.Swap(t, &trackRun, func(cfg clitrack.Config, execute func() error, exitCode func(error) int) {
+		if cfg.PID != "" || cfg.UID != "" || cfg.Username != "" {
+			t.Fatalf("opted-out tracker config = %#v", cfg)
+		}
+		if err := execute(); err != nil {
+			t.Fatalf("opted-out execution failed: %v", err)
+		}
+		if code := exitCode(nil); code != 0 {
+			t.Fatalf("opted-out exit code = %d, want 0", code)
+		}
+	})
+
 	main()
-	if !called || code != 0 {
-		t.Fatalf("main exit = called %v, code %d", called, code)
+}
+
+func TestCrossPlatformCoverageTrackerPayloadUsesReviewedFieldWhitelist(t *testing.T) {
+	testseam.Protect(t, &os.Args)
+	os.Args = []string{"dws", "sheet", "read", "--access-token", "must-not-leak"}
+	t.Setenv("SHELL", "/bin/zsh")
+	t.Setenv("TERM_SESSION_ID", "stable-session")
+	t.Setenv("TMUX_PANE", "%42")
+	t.Setenv("LANG", "zh_CN.UTF-8")
+	t.Setenv("LC_ALL", "zh_CN.UTF-8")
+	t.Chdir(t.TempDir())
+
+	requestBody := make(chan []byte, 1)
+	server := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, req *http.Request) {
+		body, _ := io.ReadAll(req.Body)
+		requestBody <- body
+		w.WriteHeader(http.StatusNoContent)
+	}))
+	defer server.Close()
+
+	commandPath := "sheet read"
+	errorMessage := ""
+	cfg := trackerConfig(app.TelemetryIdentity{UserID: "user-1", UserName: "Alice", CorpID: "corp-1"}, &commandPath, &errorMessage)
+	cfg.Endpoint = server.URL
+	cfg.FlushTimeout = time.Second
+	clitrack.New(cfg).Run(func() error { return nil }, nil)
+
+	var body []byte
+	select {
+	case body = <-requestBody:
+	case <-time.After(time.Second):
+		t.Fatal("timed out waiting for telemetry request")
+	}
+	var envelope map[string]string
+	if err := json.Unmarshal(body, &envelope); err != nil {
+		t.Fatalf("decode telemetry request %q: %v", body, err)
+	}
+	decoded, err := url.QueryUnescape(envelope["gokey"])
+	if err != nil {
+		t.Fatalf("decode gokey: %v", err)
+	}
+	globalFields, err := url.ParseQuery(decoded)
+	if err != nil {
+		t.Fatalf("parse global telemetry fields: %v", err)
+	}
+	eventFields, err := url.ParseQuery(globalFields.Get("msg"))
+	if err != nil {
+		t.Fatalf("parse event telemetry fields: %v", err)
+	}
+
+	assertTelemetryKeys(t, globalFields, []string{"app_name", "app_version", "env", "msg", "pid", "platform", "uid", "username", "version"})
+	assertTelemetryKeys(t, eventFields, []string{"c1", "c10", "c3", "c4", "c9", "p1", "p4", "ts", "type"})
+	for key, want := range map[string]string{
+		"app_name": "dws", "app_version": app.RawVersion(), "env": "prod", "pid": "wcCRwZ",
+		"platform": "cli", "uid": "user-1", "username": "Alice", "version": app.RawVersion(),
+	} {
+		if got := globalFields.Get(key); got != want {
+			t.Fatalf("global telemetry field %s = %q, want %q", key, got, want)
+		}
+	}
+	for key, want := range map[string]string{
+		"type": "event", "p1": "cli.exec", "p4": "SYS", "c1": "dws", "c3": "0", "c9": "sheet read", "c10": "corp-1",
+	} {
+		if got := eventFields.Get(key); got != want {
+			t.Fatalf("event telemetry field %s = %q, want %q", key, got, want)
+		}
+	}
+	for _, key := range []string{"device_id", "ext", "os", "os_version", "pv_id", "sdk_version", "sid", "timezone_offset"} {
+		if globalFields.Has(key) {
+			t.Fatalf("global telemetry leaked %s: %q", key, decoded)
+		}
+	}
+	for _, key := range []string{"c2", "c5", "c6", "c7", "c8"} {
+		if eventFields.Has(key) {
+			t.Fatalf("event telemetry leaked %s: %q", key, globalFields.Get("msg"))
+		}
+	}
+}
+
+func assertTelemetryKeys(t *testing.T, fields url.Values, want []string) {
+	t.Helper()
+	got := make([]string, 0, len(fields))
+	for key := range fields {
+		got = append(got, key)
+	}
+	sort.Strings(got)
+	if !slices.Equal(got, want) {
+		t.Fatalf("telemetry keys = %v, want %v", got, want)
 	}
 }

docs/cli-interface-flag-migrations.md

@@ -41,7 +41,7 @@ PR merge-base 同时拥有快照生成器、比较器和已审批清单。门禁
 scripts/policy/interface-migrations/approved-flag-migrations-v1.json
 ```
 
-清单使用严格 JSON 解析：版本、字段名大小写、JSON 值类型、命令路径和 flag 名都必须精确；拒绝重复键、未知键、scalar `null` 与尾随 JSON 值，`reason` 不能为空；禁止 `*`、`?`、前缀规则或其他 wildcard。当前清单登记了 IM ID rename 的 `pending` 记录；`pending` 只记录已评审计划，候选与 merge-base 仍必须精确匹配 `before`，因此本治理 PR **不授权 PR #904 或任何产品接口变化**。
+清单使用严格 JSON 解析：版本、字段名大小写、JSON 值类型、命令路径和 flag 名都必须精确；拒绝重复键、未知键、scalar `null` 与尾随 JSON 值，`reason` 不能为空；禁止 `*`、`?`、前缀规则或其他 wildcard。清单中的 `pending` 记录只记录已评审计划，并授权其精确列出的后续产品迁移；候选与 merge-base 仍必须精确匹配 `before`，不能授权同一个提交中的接口变化，也不能作为其他命令或参数的通配豁免。
 
 ## 两阶段迁移与回执清理
 

docs/flag-help-schema-homology.md

@@ -92,6 +92,7 @@ command/Leaf 不再写 `dws.schema.risk`；SafetySpec 走类型化 Final 载荷
 | `Required` / `MarkRequired` | 非空校验 / cobra 硬必填 | 是（`required`） |
 | `RequiredHint`, `Aliases`, `EnvVar` | 校验提示、隐藏别名、环境回退 | 否（执行细节；别名不上主 parameter 表） |
 | `ArgDefault`, `Bind`, `OmitEmpty`, `Trim`, `Transform` | toolArgs 装配语义 | 否（载荷细节；`Bind` 可进 property 映射，但不另造 flag） |
+| `Input` | 额外取值来源：`@path` 读文件 / `-` 读 stdin，在 required/enum/约束/`Validate` 之前原地解析 | 否（今日：能力由作者写进 `Usage` / `SchemaDescription` 文案，是已声明事实而非推断；不另造 flag。目标形态收敛为类型化投影字段，见 RFC §5.3） |
 
 #### 1.2.2 编排 / 执行字段（不算声明）
 

docs/rfc-command-framework-convergence.md

@@ -292,7 +292,7 @@ Definition（仅声明；不可编译）
 
 下列字段**是**框架声明面（经 `corecmd.New` 生效并嵌入 `dws.schema.*`）：
 
-- `Flags`（含 Name/Kind/Default/Required/MarkRequired/Usage 等注册面）
+- `Flags`（含 Name/Kind/Default/Required/MarkRequired/Usage 等注册面；`Input` 是取值来源声明，经 `corecmd.New` 生效但**不**嵌入 `dws.schema.*`，能力靠 `Usage` 文案声明，见 §5.3）
 - `Constraints`
 - **非空** `Risk`（空值 = 运行时当只读确认，且**不**嵌入 `dws.schema.risk`）
 - `ConstParams`（载荷声明；不上用户 flag 表）
@@ -360,6 +360,7 @@ Definition（仅声明；不可编译）
 | | `idempotency` | 评审源（或未来 Contract） | reviewed metadata | 今日非框架声明；不得推断 |
 | | `effect_source` / provenance | 组装派生物 | resolver 写入 `FieldProvenance` | 派生，不手写 |
 | **DryRun** | `preview_kind`, `remote_reads` | 评审源 | `schema_dry_run_capabilities`（正能力声明） | 否；无条目 ≠ 推断「不支持」之外的假能力 |
+| **Wait** | `mode`（`poll`/`event`/`auto`）, `poll_command`, `status_query`, `terminal`（状态→success/failure）, `pending_values`, `event_key`/`match_field`/`resource_query`（event/auto）, `default_timeout_secs` | **声明**（`ContractDecl.Wait` 正能力声明，且必须搭配 ResultInvoke dispatcher + 按模式的 hook：poll↔`WaitPoll`、event↔`WaitEvents`、auto↔两者，构造期配对校验，多余 hook 同样拒绝） | 声明后注册 `--wait`/`--wait-timeout`（框架 flag，不进 toolArgs）；Schema 投影 `wait` 键；auto = 事件优先、流终止/订阅失败回退轮询，一个 deadline 覆盖两阶段并传入 `WaitPoll`/`WaitEvents`（及 `Command().Context()`）；仅 pending 初始结果进入等待，success/failure/partial 原样返回 | 否；未声明命令传 `--wait` = unknown flag。终态失败经统一信封 `error.type: "wait"`（rc=8），超时保持 pending + `meta.operation.timed_out`（rc=0）；轮询间/轮询中/事件消费中超时一律按 pending 关闭 |
 | **Interface** | `interface_mode`, `interface_ref`, `availability`, `reason` | 评审源 | MCP meta + agent metadata 解析 | 否；与 CLI Identity 分离 |
 | **Selection** | `agent_summary`, `use_when`, `avoid_when`, `examples`, `prerequisites`, `tips`, `workflow_refs`, … | 声明（`ContractDecl.Selection` / `ProductDecl`） | `ContractDecl` / `ProductDecl`（`schema_hints/` 已退役） | 可声明；声明载荷**不得携带** `Reviewed`（旧路径专用），携带即组装报错 |
 | **FieldProvenance** | 各字段 winner / candidates | 组装派生物 | Schema 组装器 | 派生；须与 delivered value 一致 |
@@ -673,6 +674,52 @@ func (k Key[T]) Declare(opts ...FlagOption[T]) FlagSpec
 - 构造时拒绝 `InputSourceInvalid`。
 - 当前没有任何 Shortcut 或 Leaf 声明 `Input`，因此 M1 增加能力且零上线表面变化。让现有命令采用它属于 §9 下的用户可见变更。
 
+`Input` 的框架能力今日已在 `corecmd` 落地（声明即执行的过渡形态，语义与上文目标一致），使用指南：
+
+**今日声明形态**：`FlagSpec.Input []string`，源常量 `corecmd.InputFile`（`"file"`）/ `corecmd.InputStdin`（`"stdin"`）。`helpers.LeafFlag` 是 `corecmd.FlagSpec` 别名，直接可用；`shortcut.Flag.Input` 同形声明，经 `FromShortcut` 映射到 `FlagSpec`。
+
+```go
+// LeafSpec / helpers
+Flags: []helpers.LeafFlag{
+    {
+        Name:  "content",
+        Usage: "文档内容（支持 @文件路径 或 - 读 stdin）",
+        Bind:  "content",
+        Input: []string{corecmd.InputFile, corecmd.InputStdin},
+    },
+}
+
+// shortcut
+Flags: []shortcut.Flag{
+    {Name: "markdown", Desc: "Markdown 内容（支持 @文件路径 或 -）",
+     Input: []string{"file", "stdin"}},
+}
+```
+
+**运行时语义**（`resolveInputFlags`，在 `runDeclaredPreflight` 内、required/enum/约束/Validate 之前执行，原地改写 cobra flag 值）：
+
+- `--flag @path`：文件内容替换取值；`--flag -`：stdin 内容替换取值。
+- `--flag @@value`：转义为字面 `@value`，不做来源解析。
+- 只解析显式 CLI token（主名或别名）；EnvVar 回落与注册默认值透传不解析。
+- 内容前置剥离 UTF-8 BOM；`Trim` 等既有语义照常作用于解析后的值。
+- 读取失败、源不支持、`@` 后空路径都是类型化校验错误（退出码 3）；同时声明两种源而文件读取失败时附 stdin 引导 hint。
+
+**作者守则**：
+
+- 声明即全部能力：required/enum/约束/Validate 校验的已是解析后的真实内容，`Execute`/`Invoke` 无需任何额外代码。
+- `Usage`/`Desc` 必须写明支持 `@路径`/`-`；框架不自动改写 help 文案，今日也不向 Schema 投影（新增投影字段须先过 homology 评审，避免 catalog drift）。
+- `user_required` 确认的写命令若声明 `InputStdin`：stdin 在校验阶段被消费，交互确认将 fail-closed 为 `confirmation_required`，此类调用必须显式 `--yes`（或 `--dry-run`）。
+- **声明前先确认取值空间不会被前缀吃掉**：声明 `InputFile` 后，任何以 `@` 开头的合法值都会被当成文件路径（本产品尤其常见的是 at 提及类取值，如 `--at-user @zhangsan` 会报读取文件失败），用户只能改用 `@@` 转义；声明 `InputStdin` 后字面值 `-` 不可达（与 curl 等约定一致）。若该 flag 的正常取值可能命中这两种形态，就不要声明对应来源。
+- 声明在构造期校验（fail-closed panic）：仅限 `KindString`；源值必须是 `file`/`stdin` 且不重复。
+
+**今日实现与目标形态的差异**（迁移到本节目标 `FlagSpec` 时收敛）：
+
+| 维度 | 今日 | 目标 |
+|---|---|---|
+| 源类型 | `[]string` 常量 | 类型化 `InputSource` |
+| 路径边界 | 直接本地文件 IO | 复用 §5.5.2 本地文件 effect 边界 |
+| Schema 投影 | 无（靠作者在 Usage 声明） | 声明即最终源，随 Catalog 透传 |
+
 核心 FlagSpec 故意没有：
 
 - `Bind`；

docs/shortcut-public-catalog.json

@@ -1,6 +1,6 @@
 {
-  "generated_at": "2026-08-12T00:10:44.511794",
-  "count": 399,
+  "generated_at": "2026-08-18T17:38:50.904696",
+  "count": 436,
   "results": [
     {
       "suite": "semantic",
@@ -489,7 +489,7 @@
       "risk": "read",
       "status": "reviewed_available",
       "disposition": "semantic_adapter",
-      "semantic_delta": "统一 ID、筛选、关键词和分页查询，并投影稳定的 records/count/cursor 结果。",
+      "semantic_delta": "统一 ID、筛选、关键词、字段投影和分页查询；fieldIds 只返回用户要求的列，并投影稳定的 records/count/cursor 结果。",
       "availability": "available"
     },
     {
@@ -692,6 +692,16 @@
       "semantic_delta": "按当前父目录语义调整文件夹展示顺序。",
       "availability": "available"
     },
+    {
+      "suite": "semantic",
+      "service": "aitable",
+      "command": "+table-bootstrap",
+      "risk": "write",
+      "status": "reviewed_available",
+      "disposition": "primary_smart",
+      "semantic_delta": "在已有 Base 中声明式创建一张表和字段，字段自动分片，逐层读回验证并发布可恢复检查点。",
+      "availability": "available"
+    },
     {
       "suite": "semantic",
       "service": "aitable",
@@ -1056,144 +1066,274 @@
       "status": "real-ok"
     },
     {
-      "suite": "read",
+      "suite": "semantic",
       "service": "calendar",
       "command": "+agenda",
       "risk": "read",
-      "status": "real-ok"
+      "status": "reviewed_available",
+      "disposition": "semantic_adapter",
+      "semantic_delta": "严格验证 result.events 与分页证据，稳定投影 id 为 eventId；显式空数组才是空日程。",
+      "availability": "available"
     },
     {
-      "suite": "read",
+      "suite": "semantic",
       "service": "calendar",
       "command": "+attendee-list",
       "risk": "read",
-      "status": "real-ok"
+      "status": "reviewed_available",
+      "disposition": "semantic_adapter",
+      "semantic_delta": "严格验证参会人数组并投影显示名、响应状态和 self；真实后端不返回稳定参会人 userId，因此不伪造身份字段。",
+      "availability": "available"
     },
     {
-      "suite": "write",
+      "suite": "semantic",
       "service": "calendar",
       "command": "+book",
       "risk": "write",
-      "status": "real-ok"
+      "status": "reviewed_available",
+      "disposition": "primary_smart",
+      "semantic_delta": "按参会人姓名解析后创建日程，并在多步骤路径提供失败回滚与详情读回。",
+      "availability": "available"
     },
     {
-      "suite": "read",
+      "suite": "semantic",
       "service": "calendar",
       "command": "+book-list",
       "risk": "read",
-      "status": "real-ok"
+      "status": "reviewed_available",
+      "disposition": "semantic_adapter",
+      "semantic_delta": "严格验证日历本数组并稳定投影 calendarId，缺失数组不再误报为空。",
+      "availability": "available"
     },
     {
-      "suite": "read",
+      "suite": "semantic",
       "service": "calendar",
       "command": "+book-search",
       "risk": "read",
-      "status": "real-ok"
+      "status": "reviewed_available",
+      "disposition": "semantic_adapter",
+      "semantic_delta": "按日历本名称搜索并严格区分零命中与协议错误。",
+      "availability": "available"
     },
     {
-      "suite": "write",
+      "suite": "semantic",
       "service": "calendar",
       "command": "+cancel-event",
       "risk": "high-risk-write",
-      "status": "real-ok"
+      "status": "reviewed_available",
+      "disposition": "primary_smart",
+      "semantic_delta": "删除前读取并展示目标，经高风险确认后取消指定日程。",
+      "availability": "available"
     },
     {
-      "suite": "read",
+      "suite": "semantic",
       "service": "calendar",
       "command": "+conflicts",
       "risk": "read",
-      "status": "real-ok"
+      "status": "reviewed_available",
+      "disposition": "primary_smart",
+      "semantic_delta": "在指定范围分析重叠日程并输出冲突对，避免人工逐项比对。",
+      "availability": "available"
     },
     {
-      "suite": "read",
+      "suite": "semantic",
+      "service": "calendar",
+      "command": "+create",
+      "risk": "write",
+      "status": "reviewed_available",
+      "disposition": "primary_smart",
+      "semantic_delta": "创建响应必须给出 eventId/id，且标题和起止时间通过详情读回后才成功。",
+      "availability": "available"
+    },
+    {
+      "suite": "semantic",
       "service": "calendar",
       "command": "+free",
       "risk": "read",
-      "status": "real-ok"
+      "status": "reviewed_available",
+      "disposition": "primary_smart",
+      "semantic_delta": "按姓名解析参与者后查询闲忙，省去手工 userId 解析。",
+      "availability": "available"
     },
     {
-      "suite": "read",
+      "suite": "semantic",
       "service": "calendar",
       "command": "+free-slots",
       "risk": "read",
-      "status": "real-ok"
+      "status": "reviewed_available",
+      "disposition": "primary_smart",
+      "semantic_delta": "从已有日程计算工作时段内空档，并显式输出满足时长的候选区间。",
+      "availability": "available"
     },
     {
-      "suite": "read",
+      "suite": "semantic",
       "service": "calendar",
       "command": "+freebusy",
       "risk": "read",
-      "status": "real-ok"
+      "status": "reviewed_available",
+      "disposition": "semantic_adapter",
+      "semantic_delta": "统一按用户或会议室 ID 查询指定范围闲忙，并声明至少一类目标约束。",
+      "availability": "available"
     },
     {
-      "suite": "write",
+      "suite": "semantic",
+      "service": "calendar",
+      "command": "+get",
+      "risk": "read",
+      "status": "reviewed_available",
+      "disposition": "schema_leaf",
+      "semantic_delta": "严格读取日程对象并把稳定 id 规范化为 eventId，拒绝空响应和对象漂移。",
+      "availability": "available"
+    },
+    {
+      "suite": "semantic",
       "service": "calendar",
       "command": "+invite",
       "risk": "write",
-      "status": "real-ok"
+      "status": "reviewed_available",
+      "disposition": "primary_smart",
+      "semantic_delta": "按姓名唯一解析用户后邀请到已有日程。",
+      "availability": "available"
     },
     {
-      "suite": "read",
+      "suite": "semantic",
       "service": "calendar",
       "command": "+my-free",
       "risk": "read",
-      "status": "real-ok"
+      "status": "reviewed_available",
+      "disposition": "primary_smart",
+      "semantic_delta": "以当前用户身份查询一个时间段是否空闲。",
+      "availability": "available"
     },
     {
-      "suite": "read",
+      "suite": "semantic",
       "service": "calendar",
       "command": "+next-event",
       "risk": "read",
-      "status": "real-ok"
+      "status": "reviewed_available",
+      "disposition": "primary_smart",
+      "semantic_delta": "在未来时间窗内定位最近的下一场日程。",
+      "availability": "available"
     },
     {
-      "suite": "write",
+      "suite": "semantic",
       "service": "calendar",
       "command": "+reschedule",
       "risk": "write",
-      "status": "real-ok"
+      "status": "reviewed_available",
+      "disposition": "primary_smart",
+      "semantic_delta": "修改前读取目标日程，再更新完整起止时间。",
+      "availability": "available"
     },
     {
-      "suite": "read",
+      "suite": "semantic",
+      "service": "calendar",
+      "command": "+room-find",
+      "risk": "read",
+      "status": "reviewed_available",
+      "disposition": "semantic_adapter",
+      "semantic_delta": "按单个未来时间段严格查询可用会议室，公开真实 page/pageSize/hasMore 证据和后端筛选边界。",
+      "availability": "available"
+    },
+    {
+      "suite": "semantic",
       "service": "calendar",
       "command": "+room-groups",
       "risk": "read",
-      "status": "real-ok"
+      "status": "reviewed_available",
+      "disposition": "semantic_adapter",
+      "semantic_delta": "严格读取 result.groupList，避免会议室分组被响应 wrapper 漂移静默清空。",
+      "availability": "available"
     },
     {
-      "suite": "read",
+      "suite": "semantic",
       "service": "calendar",
       "command": "+room-search",
       "risk": "read",
-      "status": "real-ok"
+      "status": "reviewed_available",
+      "disposition": "semantic_adapter",
+      "semantic_delta": "按名称查询会议室但不声称检查时间可用性；缺失会议室数组直接失败。",
+      "availability": "available"
     },
     {
-      "suite": "read",
+      "suite": "semantic",
+      "service": "calendar",
+      "command": "+rsvp",
+      "risk": "write",
+      "status": "reviewed_available",
+      "disposition": "semantic_adapter",
+      "semantic_delta": "把直白 RSVP 动作映射为 responseStatus，拒绝空 ack，并在写后读取日程终态。",
+      "availability": "available"
+    },
+    {
+      "suite": "semantic",
+      "service": "calendar",
+      "command": "+search-event",
+      "risk": "read",
+      "status": "reviewed_available",
+      "disposition": "semantic_adapter",
+      "semantic_delta": "在服务端日程当前页按标题、描述和地点过滤，同时保留 hasMore/nextCursor，绝不把单页零命中当全局未找到。",
+      "availability": "available"
+    },
+    {
+      "suite": "semantic",
       "service": "calendar",
       "command": "+suggest-time",
       "risk": "read",
-      "status": "real-ok"
+      "status": "reviewed_available",
+      "disposition": "primary_smart",
+      "semantic_delta": "按姓名逐个唯一解析参与者，再查询大家共同可用的建议时间。",
+      "availability": "available"
     },
     {
-      "suite": "read",
+      "suite": "semantic",
+      "service": "calendar",
+      "command": "+suggestion",
+      "risk": "read",
+      "status": "reviewed_available",
+      "disposition": "schema_leaf",
+      "semantic_delta": "直接按 userId 查询建议时段，严格区分显式空 recommendEventTimes 与缺失或畸形响应。",
+      "availability": "available"
+    },
+    {
+      "suite": "semantic",
       "service": "calendar",
       "command": "+today",
       "risk": "read",
-      "status": "real-ok"
+      "status": "reviewed_available",
+      "disposition": "primary_smart",
+      "semantic_delta": "查询并整理今天的日程视图。",
+      "availability": "available"
     },
     {
-      "suite": "read",
+      "suite": "semantic",
       "service": "calendar",
       "command": "+tomorrow",
       "risk": "read",
-      "status": "real-ok"
+      "status": "reviewed_available",
+      "disposition": "primary_smart",
+      "semantic_delta": "查询并整理明天的日程视图。",
+      "availability": "available"
     },
     {
-      "suite": "read",
+      "suite": "semantic",
+      "service": "calendar",
+      "command": "+update",
+      "risk": "write",
+      "status": "reviewed_available",
+      "disposition": "primary_smart",
+      "semantic_delta": "组合字段与参会人分阶段更新，明确非事务步骤并以详情和参会人列表读回验证。",
+      "availability": "available"
+    },
+    {
+      "suite": "semantic",
       "service": "calendar",
       "command": "+week",
       "risk": "read",
-      "status": "real-ok"
+      "status": "reviewed_available",
+      "disposition": "primary_smart",
+      "semantic_delta": "查询并按日期整理本周日程。",
+      "availability": "available"
     },
     {
       "suite": "semantic",
@@ -3582,88 +3722,414 @@
       "status": "real-ok"
     },
     {
-      "suite": "write",
+      "suite": "semantic",
       "service": "todo",
       "command": "+assign",
       "risk": "write",
-      "status": "real-ok"
+      "status": "reviewed_available",
+      "disposition": "primary_smart",
+      "semantic_delta": "保留历史按姓名解析后创建并指派的新任务语义，避免同名破坏性变更。",
+      "availability": "available"
     },
     {
-      "suite": "write",
+      "suite": "semantic",
       "service": "todo",
       "command": "+assign-multi",
       "risk": "write",
-      "status": "real-ok"
+      "status": "reviewed_available",
+      "disposition": "primary_smart",
+      "semantic_delta": "全部姓名唯一解析后才创建一次，并以单一统一结果输出稳定 taskId。",
+      "availability": "available"
     },
     {
-      "suite": "read",
+      "suite": "semantic",
+      "service": "todo",
+      "command": "+comment",
+      "risk": "write",
+      "status": "reviewed_available",
+      "disposition": "primary_smart",
+      "semantic_delta": "新增评论必须取得稳定 commentId，并分页读回验证内容。",
+      "availability": "available"
+    },
+    {
+      "suite": "semantic",
+      "service": "todo",
+      "command": "+complete",
+      "risk": "write",
+      "status": "reviewed_available",
+      "disposition": "primary_smart",
+      "semantic_delta": "先读当前状态避免重复写，必要时完成后再读回核验。",
+      "availability": "available"
+    },
+    {
+      "suite": "semantic",
+      "service": "todo",
+      "command": "+create",
+      "risk": "write",
+      "status": "reviewed_available",
+      "disposition": "primary_smart",
+      "semantic_delta": "创建后要求稳定 taskId，并读取 todoDetailModel 验证标题。",
+      "availability": "available"
+    },
+    {
+      "suite": "semantic",
       "service": "todo",
       "command": "+created-todos",
       "risk": "read",
-      "status": "real-ok"
+      "status": "reviewed_available",
+      "disposition": "primary_smart",
+      "semantic_delta": "遍历创建者角色的全部分页后投影。",
+      "availability": "available"
     },
     {
-      "suite": "read",
+      "suite": "semantic",
+      "service": "todo",
+      "command": "+due-today",
+      "risk": "read",
+      "status": "reviewed_available",
+      "disposition": "primary_smart",
+      "semantic_delta": "按本地日历日构造服务端截止窗口并遍历全部分页。",
+      "availability": "available"
+    },
+    {
+      "suite": "semantic",
       "service": "todo",
       "command": "+get",
       "risk": "read",
-      "status": "real-ok"
+      "status": "reviewed_available",
+      "disposition": "semantic_adapter",
+      "semantic_delta": "严格绑定 todoDetailModel.taskId，缺失或错绑均失败。",
+      "availability": "available"
     },
     {
-      "suite": "read",
+      "suite": "semantic",
       "service": "todo",
       "command": "+get-my-tasks",
       "risk": "read",
-      "status": "real-ok"
+      "status": "reviewed_available",
+      "disposition": "semantic_adapter",
+      "semantic_delta": "严格识别 todoCards/hasMore/page/size；显式空数组才是空结果，并支持有界全量分页。",
+      "availability": "available"
     },
     {
-      "suite": "read",
+      "suite": "semantic",
+      "service": "todo",
+      "command": "+get-related-tasks",
+      "risk": "read",
+      "status": "reviewed_available",
+      "disposition": "primary_smart",
+      "semantic_delta": "遍历与我相关三种角色的全部分页并按 taskId 去重。",
+      "availability": "available"
+    },
+    {
+      "suite": "semantic",
       "service": "todo",
       "command": "+list-attachment",
       "risk": "read",
-      "status": "real-ok"
+      "status": "reviewed_available",
+      "disposition": "semantic_adapter",
+      "semantic_delta": "严格区分显式空附件与缺失容器，并要求每项稳定附件 ID。",
+      "availability": "available"
     },
     {
-      "suite": "read",
+      "suite": "semantic",
       "service": "todo",
       "command": "+list-comment",
       "risk": "read",
-      "status": "real-ok"
+      "status": "reviewed_available",
+      "disposition": "semantic_adapter",
+      "semantic_delta": "严格区分显式空评论与缺失容器，并要求每项稳定 commentId。",
+      "availability": "available"
     },
     {
-      "suite": "read",
+      "suite": "semantic",
       "service": "todo",
       "command": "+list-sub",
       "risk": "read",
-      "status": "real-ok"
+      "status": "reviewed_available",
+      "disposition": "semantic_adapter",
+      "semantic_delta": "严格校验子待办容器、对象元素和稳定 taskId。",
+      "availability": "available"
     },
     {
-      "suite": "read",
+      "suite": "semantic",
       "service": "todo",
       "command": "+overdue",
       "risk": "read",
-      "status": "real-ok"
+      "status": "reviewed_available",
+      "disposition": "primary_smart",
+      "semantic_delta": "遍历执行者角色的全部分页并按当前时间筛选逾期项。",
+      "availability": "available"
     },
     {
-      "suite": "write",
+      "suite": "semantic",
+      "service": "todo",
+      "command": "+related-tasks",
+      "risk": "read",
+      "status": "reviewed_available",
+      "disposition": "alias_internal",
+      "semantic_delta": "保留旧命令拼写，新的 Agent 路由统一使用 +get-related-tasks。",
+      "availability": "available"
+    },
+    {
+      "suite": "semantic",
       "service": "todo",
       "command": "+remind",
       "risk": "write",
-      "status": "real-ok"
+      "status": "reviewed_available",
+      "disposition": "primary_smart",
+      "semantic_delta": "解析当前用户后创建给自己的待办；该命令不是独立提醒规则。",
+      "availability": "available"
     },
     {
-      "suite": "write",
+      "suite": "semantic",
+      "service": "todo",
+      "command": "+reminder",
+      "risk": "write",
+      "status": "reviewed_available",
+      "disposition": "semantic_adapter",
+      "semantic_delta": "严格验证提醒参数和 success=true 终端回执；上游不能读回规则，固定 verified=false。",
+      "availability": "available"
+    },
+    {
+      "suite": "semantic",
+      "service": "todo",
+      "command": "+reopen",
+      "risk": "write",
+      "status": "reviewed_available",
+      "disposition": "primary_smart",
+      "semantic_delta": "先读当前状态避免重复写，必要时重开后再读回核验。",
+      "availability": "available"
+    },
+    {
+      "suite": "semantic",
+      "service": "todo",
+      "command": "+search",
+      "risk": "read",
+      "status": "reviewed_available",
+      "disposition": "primary_smart",
+      "semantic_delta": "遍历全部 hasMore 分页并按标题匹配；达到页上限仍未耗尽时失败。",
+      "availability": "available"
+    },
+    {
+      "suite": "semantic",
       "service": "todo",
       "command": "+todo-done",
       "risk": "write",
-      "status": "real-ok"
+      "status": "reviewed_available",
+      "disposition": "primary_smart",
+      "semantic_delta": "跨全部分页按标题唯一定位后完成任务。",
+      "availability": "available"
     },
     {
-      "suite": "read",
+      "suite": "semantic",
+      "service": "todo",
+      "command": "+update",
+      "risk": "write",
+      "status": "reviewed_available",
+      "disposition": "primary_smart",
+      "semantic_delta": "更新指定字段后读取详情逐字段核验。",
+      "availability": "available"
+    },
+    {
+      "suite": "semantic",
+      "service": "wiki",
+      "command": "+delete-space",
+      "risk": "high-risk-write",
+      "status": "reviewed_available",
+      "disposition": "semantic_adapter",
+      "semantic_delta": "删除前读取影响目标，经高风险确认后只接受 success=true 终态。",
+      "availability": "available"
+    },
+    {
+      "suite": "semantic",
+      "service": "wiki",
+      "command": "+feed-list",
+      "risk": "read",
+      "status": "reviewed_available",
+      "disposition": "semantic_adapter",
+      "semantic_delta": "补充知识库协作动态查询，严格验证 feeds 并保留游标。",
+      "availability": "available"
+    },
+    {
+      "suite": "semantic",
+      "service": "wiki",
+      "command": "+member-add",
+      "risk": "write",
+      "status": "reviewed_available",
+      "disposition": "semantic_adapter",
+      "semantic_delta": "支持 1-30 个 userId 与四类角色；严格要求写接口终态，并明确后端无法提供精确成员读回。",
+      "availability": "available"
+    },
+    {
+      "suite": "semantic",
+      "service": "wiki",
+      "command": "+member-list",
+      "risk": "read",
+      "status": "reviewed_available",
+      "disposition": "semantic_adapter",
+      "semantic_delta": "严格验证成员数组并公开真实单次上限 50；后端无游标时不伪造 page-all。",
+      "availability": "available"
+    },
+    {
+      "suite": "semantic",
+      "service": "wiki",
+      "command": "+member-remove",
+      "risk": "write",
+      "status": "reviewed_available",
+      "disposition": "semantic_adapter",
+      "semantic_delta": "支持批量 userId 移除；严格要求写接口终态，并明确后端无法提供精确成员读回。",
+      "availability": "available"
+    },
+    {
+      "suite": "semantic",
+      "service": "wiki",
+      "command": "+member-update",
+      "risk": "write",
+      "status": "reviewed_available",
+      "disposition": "semantic_adapter",
+      "semantic_delta": "补充成员角色更新；严格要求写接口终态，并明确后端无法提供精确成员读回。",
+      "availability": "available"
+    },
+    {
+      "suite": "semantic",
+      "service": "wiki",
+      "command": "+move",
+      "risk": "write",
+      "status": "reviewed_available",
+      "disposition": "primary_smart",
+      "semantic_delta": "同一入口覆盖 Wiki 内移动和我的文档在线节点入 Wiki，并校验目标 workspace/folder。",
+      "availability": "available"
+    },
+    {
+      "suite": "semantic",
+      "service": "wiki",
+      "command": "+move-to-drive",
+      "risk": "write",
+      "status": "reviewed_available",
+      "disposition": "primary_smart",
+      "semantic_delta": "同步移动 Wiki 节点到我的文档并读回验证 workspace 变化，免去异步任务轮询。",
+      "availability": "available"
+    },
+    {
+      "suite": "semantic",
+      "service": "wiki",
+      "command": "+node-copy",
+      "risk": "high-risk-write",
+      "status": "reviewed_available",
+      "disposition": "primary_smart",
+      "semantic_delta": "高风险确认后复制，必须取得新 nodeId 并读回副本，避免空响应被当作成功。",
+      "availability": "available"
+    },
+    {
+      "suite": "semantic",
+      "service": "wiki",
+      "command": "+node-create",
+      "risk": "write",
+      "status": "reviewed_available",
+      "disposition": "primary_smart",
+      "semantic_delta": "支持七种 Wiki 节点类型，创建后要求 nodeId 并读取元数据验证。",
+      "availability": "available"
+    },
+    {
+      "suite": "semantic",
+      "service": "wiki",
+      "command": "+node-delete",
+      "risk": "high-risk-write",
+      "status": "reviewed_available",
+      "disposition": "semantic_adapter",
+      "semantic_delta": "删除前读取并核对 workspace，经高风险确认后要求 success=true 终态证据。",
+      "availability": "available"
+    },
+    {
+      "suite": "semantic",
+      "service": "wiki",
+      "command": "+node-get",
+      "risk": "read",
+      "status": "reviewed_available",
+      "disposition": "schema_leaf",
+      "semantic_delta": "统一节点 ID 或在线文档 URL 的元数据读取，补充文档域属性视角。",
+      "availability": "available"
+    },
+    {
+      "suite": "semantic",
+      "service": "wiki",
+      "command": "+node-list",
+      "risk": "read",
+      "status": "reviewed_available",
+      "disposition": "semantic_adapter",
+      "semantic_delta": "严格区分显式空目录与假空成功，稳定投影节点并保留 nextCursor/hasMore。",
+      "availability": "available"
+    },
+    {
+      "suite": "semantic",
+      "service": "wiki",
+      "command": "+node-search",
+      "risk": "read",
+      "status": "reviewed_available",
+      "disposition": "semantic_adapter",
+      "semantic_delta": "补充库内关键词和扩展名搜索，并拒绝假空结果。",
+      "availability": "available"
+    },
+    {
+      "suite": "semantic",
+      "service": "wiki",
+      "command": "+resolve-space",
+      "risk": "read",
+      "status": "reviewed_available",
+      "disposition": "primary_smart",
+      "semantic_delta": "把关键词搜索收敛为唯一 workspaceId；零命中与多命中显式分流，绝不猜测。",
+      "availability": "available"
+    },
+    {
+      "suite": "semantic",
+      "service": "wiki",
+      "command": "+space-create",
+      "risk": "write",
+      "status": "reviewed_available",
+      "disposition": "primary_smart",
+      "semantic_delta": "创建后要求 workspaceId 并通过空间详情读回验证真实落库。",
+      "availability": "available"
+    },
+    {
+      "suite": "semantic",
+      "service": "wiki",
+      "command": "+space-get",
+      "risk": "read",
+      "status": "reviewed_available",
+      "disposition": "schema_leaf",
+      "semantic_delta": "补充知识库详情入口，并要求 workspaceId 业务证据。",
+      "availability": "available"
+    },
+    {
+      "suite": "semantic",
+      "service": "wiki",
+      "command": "+space-list",
+      "risk": "read",
+      "status": "reviewed_available",
+      "disposition": "semantic_adapter",
+      "semantic_delta": "严格区分显式空知识库列表与缺失、畸形或内部错误响应，并保留真实分页证据。",
+      "availability": "available"
+    },
+    {
+      "suite": "semantic",
       "service": "wiki",
       "command": "+space-search",
       "risk": "read",
-      "status": "real-ok"
+      "status": "reviewed_available",
+      "disposition": "semantic_adapter",
+      "semantic_delta": "按关键词搜索知识库并拒绝把缺失业务数组误报为零命中。",
+      "availability": "available"
+    },
+    {
+      "suite": "semantic",
+      "service": "wiki",
+      "command": "+wiki-new-doc",
+      "risk": "write",
+      "status": "reviewed_available",
+      "disposition": "primary_smart",
+      "semantic_delta": "按空间名精确唯一解析、创建在线文档并读回验证；零命中和歧义均显式失败。",
+      "availability": "available"
     }
   ]
 }

docs/shortcut-report.html

@@ -102,7 +102,7 @@
       <tr><td><code>minutes +latest-minutes</code></td><td>列妙记→取最新一条详情</td><td class="c ok">编译/挂载</td></tr>
       <tr><td><code>minutes +action-items</code></td><td>列妙记→取最新→取其待办</td><td class="c ok">编译/挂载</td></tr>
       <tr><td><code>wiki +wiki-new-doc --space &lt;名&gt;</code></td><td>按名搜知识空间→建文档（跨 doc server 路由）</td><td class="c ok">编译/挂载</td></tr>
-      <tr><td><code>doc +doc-append --doc --text</code></td><td>文档末尾追加文本</td><td class="c ok">编译/挂载</td></tr>
+      <tr><td><code>doc +doc-append --doc --content</code></td><td>文档末尾追加文本</td><td class="c ok">编译/挂载</td></tr>
       <tr><td><code>doc +share-doc --to &lt;名&gt; --url</code></td><td>解析人→把文档链接私信 TA（跨服务）</td><td class="c ok">编译/挂载</td></tr>
     </tbody>
   </table>

docs/shortcut-report.md

@@ -56,13 +56,13 @@
 
 | shortcut | 多步/智能逻辑 | 验证 |
 |----------|--------------|------|
-| `chat +dm --to <姓名> --text` | 搜人→解析唯一 userId→发单聊；多人消歧 | ✅ dry-run 真机 |
+| `chat +dm --to <姓名> --content` | 搜人→解析唯一 userId→发单聊；多人消歧 | ✅ dry-run 真机 |
 | `contact +lookup --name <姓名>` | 搜人→解析 userId→取完整资料 | ✅ **真机端到端** |
 | `todo +assign --to <姓名> --task` | 解析人→建待办并把 TA 设为执行人 | ✅ dry-run 真机 |
-| `chat +send-to-group --group <群名> --text` | 按群名搜群(search_groups)→消歧→发消息 | ✅ 编译/挂载 |
+| `chat +send-to-group --group <群名> --content` | 按群名搜群(search_groups)→消歧→发消息 | ✅ 编译/挂载 |
 | `calendar +book --title --start --end [--with <姓名CSV>]` | 建日程→按名加参与者→**失败回滚删日程**（对标 lark `calendar +create`） | ✅ dry-run 真机 |
 | `calendar +free --who <姓名> --start --end` | 解析人→查其时段忙闲 | ✅ **真机端到端**（解析 202397→查忙闲） |
-| `chat +broadcast --to <姓名CSV> --text` | 多名逐一解析→群发单聊，失败汇总不中断 | ✅ 编译/挂载 |
+| `chat +broadcast --to <姓名CSV> --content` | 多名逐一解析→群发单聊，失败汇总不中断 | ✅ 编译/挂载 |
 | `minutes +latest-minutes` | 列妙记→取最新一条详情 | ✅ 编译/挂载 |
 | `chat +group-members --group <群名>` | 按群名搜群→列群成员 | ✅ 编译/挂载 |
 | `contact +org --name <姓名>` | 解析人→取详情拿 deptId→查部门详情 | ✅ **真机端到端**（3 步：董鑫阳→模型算法/16人） |
@@ -76,7 +76,7 @@
 | `todo +todo-done --task <关键词>` | 列我的待办→按标题匹配→标记完成 | ✅ 编译/挂载 |
 | `calendar +reschedule --event <id>` | 查日程详情→改时间（查→改机械多步） | ✅ 编译/挂载 |
 | `wiki +wiki-new-doc --space <名>` | 按名搜知识空间→在其下建文档（跨 doc server 路由） | ✅ 编译/挂载 |
-| `doc +doc-append --doc --text` | 文档末尾追加文本（update_document append 模式） | ✅ 编译/挂载 |
+| `doc +doc-append --doc --content` | 文档末尾追加文本（update_document append 模式） | ✅ 编译/挂载 |
 | `minutes +action-items` | 列妙记→取最新→取其待办事项 | ✅ 编译/挂载 |
 | `minutes +detail --id <taskUuid>` | 一条命令聚合听记 basic/summary/keywords/transcript/todos，partial-failure 容错 | ✅ 全量测试 |
 | `minutes +replace-batch --id --pair "原文=>替换"…` | 多组批量替换文字，去重校验+逐组结果聚合 | ✅ 全量测试 |

docs/superpowers/specs/2026-08-14-oa-attachment-download-url-output-design.md

@@ -0,0 +1,55 @@
+# OA Attachment Download URL Output Design
+
+## Goal
+
+Keep the existing command and MCP request unchanged while making the returned
+OSS signed URL directly copyable from JSON output:
+
+```text
+dws oa approval attachment download-url
+```
+
+## Scope
+
+Only `oa approval attachment download-url` changes. The other OA attachment
+commands and the global JSON formatter retain their current behavior.
+
+## Design
+
+The command continues to invoke MCP server `oa`, tool
+`get_attachment_download_url`, with the same arguments. Its leaf declaration
+provides a command-specific `Call` callback that invokes the existing MCP
+dispatcher with HTML escaping disabled when the selected output format is
+JSON. This preserves literal `&` separators in `result.downloadUri` instead of
+rendering them as `\u0026`.
+
+For `raw`, `table`, and other non-JSON formats, the callback uses the existing
+escaped dispatcher behavior so their current rendering remains unchanged.
+
+The change does not alter the URL, decode or re-sign it, download the file, or
+change global JSON serialization.
+
+## Error Handling
+
+Authentication, MCP transport, gateway, PAT, and business errors continue
+through the existing dispatcher and retain their current classification and
+output behavior.
+
+## Verification
+
+Add a `TestCrossPlatformCoverage*` regression test that executes the real Cobra
+leaf in explicit JSON mode with a fake MCP result containing a signed URL. It
+must verify:
+
+- the request still targets `oa/get_attachment_download_url`;
+- the exact request arguments remain unchanged, including omission of the
+  optional boolean when the flag was not supplied;
+- stdout contains literal `&OSSAccessKeyId=` and `&Signature=`;
+- stdout contains no `\u0026` escape.
+
+The fake caller must report JSON format (or the command must be executed with
+`--format json`) so the test fails against the current escaped JSON path rather
+than accidentally exercising raw MCP text output.
+
+Run the focused OA attachment tests, format modified Go files, and rebuild the
+CLI. No commit is created.

docs/wiki-shortcut-business-review.html

@@ -0,0 +1,117 @@
+<!doctype html>
+<html lang="zh-CN">
+<head>
+  <meta charset="utf-8">
+  <meta name="viewport" content="width=device-width,initial-scale=1">
+  <title>DWS Wiki Shortcut 全景评审</title>
+  <style>
+    :root{--ink:#14213d;--muted:#5c677d;--line:#dbe4f0;--paper:#fff;--bg:#f3f7fb;--blue:#1769e0;--cyan:#00a6a6;--green:#178746;--amber:#a45b00;--red:#b42318;--shadow:0 14px 34px rgba(20,33,61,.08)}
+    *{box-sizing:border-box}body{margin:0;overflow-x:hidden;background:linear-gradient(150deg,#edf5ff 0,#f8fbff 45%,#eef8f5 100%);color:var(--ink);font:15px/1.65 -apple-system,BlinkMacSystemFont,"Segoe UI","PingFang SC",sans-serif}
+    main,.card,.two>*{min-width:0}main{width:min(1180px,calc(100% - 32px));margin:28px auto 72px}.hero,.card{background:rgba(255,255,255,.96);border:1px solid var(--line);border-radius:22px;box-shadow:var(--shadow)}
+    .hero{padding:38px;background:radial-gradient(circle at 95% 0,#dff8f3,transparent 36%),linear-gradient(135deg,#fff,#f5f9ff)}h1{font-size:34px;line-height:1.2;margin:0 0 10px}.lead{font-size:17px;color:var(--muted);max-width:900px}.meta{display:flex;gap:10px;flex-wrap:wrap;margin-top:20px}.pill{border:1px solid #cbd9ea;border-radius:999px;padding:5px 11px;background:#fff;font-size:13px}
+    .grid{display:grid;grid-template-columns:repeat(4,1fr);gap:14px;margin:18px 0}.metric{padding:20px}.metric b{display:block;font-size:31px;color:var(--blue)}.metric span{color:var(--muted)}
+    section{margin-top:22px}.card{padding:26px}h2{font-size:23px;margin:0 0 14px}h3{font-size:17px;margin:22px 0 8px}.callout{border-left:4px solid var(--blue);background:#f2f7ff;padding:14px 16px;border-radius:8px}.warn{border-color:var(--amber);background:#fff8eb}.ok{border-color:var(--green);background:#effbf4}
+    table{width:100%;border-collapse:collapse;font-size:14px}th,td{text-align:left;vertical-align:top;border-bottom:1px solid var(--line);padding:11px 9px}th{color:#41516b;background:#f7f9fc;position:sticky;top:0}code{background:#edf2f8;border-radius:5px;padding:2px 5px;color:#24466e}.tag{display:inline-block;border-radius:999px;padding:2px 8px;font-size:12px;font-weight:650;white-space:nowrap}.full{background:#e6f6ec;color:#116436}.partial{background:#fff0d5;color:#875000}.extra{background:#e8f1ff;color:#1854a5}.fixed{background:#f1eaff;color:#6338a5}
+    .toolbar{display:flex;flex-wrap:wrap;gap:10px;margin:12px 0}.toolbar input,.toolbar select{border:1px solid #bdcada;border-radius:10px;padding:9px 11px;background:#fff;min-width:min(220px,100%);max-width:100%;flex:1 1 220px}.matrix{max-height:620px;overflow:auto;border:1px solid var(--line);border-radius:12px}.two{display:grid;grid-template-columns:1fr 1fr;gap:18px}.small{color:var(--muted);font-size:13px}ul{padding-left:20px}.footer{color:var(--muted);text-align:center;margin-top:22px}@media(max-width:850px){.grid,.two{grid-template-columns:1fr 1fr}.hero{padding:25px}}@media(max-width:560px){.grid,.two{grid-template-columns:1fr}main{width:min(100% - 18px,1180px)}.card{padding:18px}h1{font-size:28px}}
+  </style>
+</head>
+<body><main>
+  <header class="hero">
+    <h1>DWS Wiki Shortcut 全景评审</h1>
+    <p class="lead">以 13 项成熟 Wiki 用户任务为基线，重新审视 DWS 的空间、成员、节点与动态能力。本次不是按命令名凑数：每个入口都要求真实业务证据，缺失数组、畸形响应、空确认或读回不一致一律失败。</p>
+    <div class="meta"><span class="pill">评审日期 2026-08-14</span><span class="pill">独立 worktree / 独立分支</span><span class="pill">真实组织数据 E2E 28/28</span><span class="pill">报告已去标识化</span></div>
+  </header>
+
+  <div class="grid">
+    <div class="card metric"><b>20</b><span>公开 Wiki Shortcuts</span></div>
+    <div class="card metric"><b>13/13</b><span>基线用户任务有对应路径</span></div>
+    <div class="card metric"><b>7</b><span>DWS 额外场景</span></div>
+    <div class="card metric"><b>20/20</b><span>真实数据能力已触达</span></div>
+  </div>
+
+  <section class="card">
+    <h2>结论先行</h2>
+    <div class="callout ok"><strong>DWS 已形成比“API 快捷别名”更完整的 Wiki 任务层。</strong> 基线中的 13 个用户任务均有对应入口；DWS 还提供空间搜索/详情/唯一解析、成员角色更新、库内节点搜索、协作动态和按空间名新建文档。创建、复制、移动等关键写能力从“请求发出”升级为“终态 + ID + 读回”成功标准。</div>
+    <div class="callout warn" style="margin-top:12px"><strong>能力边界必须诚实表达。</strong> DingTalk 成员接口不提供游标，单次真实上限是 50，因此不能实现成员 <code>--page-all</code>；成员身份只接受同组织可用的 userId，无法提供 email/open_id 等多种身份模式；节点创建也没有等价的 origin/shortcut 模式。这些差异保留为明确边界，而不是用本地循环或空结果伪装。</div>
+  </section>
+
+  <section class="card">
+    <h2>13 项基线任务逐条映射</h2>
+    <div class="matrix"><table><thead><tr><th>基线任务</th><th>DWS 主入口</th><th>结论</th><th>DWS 视角与边界</th></tr></thead><tbody>
+      <tr><td><code>+space-list</code></td><td><code>wiki +space-list</code></td><td><span class="tag full">完整对齐</span></td><td>严格空集合、游标续传、自动翻页、停滞检测；支持组织/我的知识库。</td></tr>
+      <tr><td><code>+space-create</code></td><td><code>wiki +space-create</code></td><td><span class="tag full">超过</span></td><td>公开真实 32 字符名称上限；创建后按 workspaceId 读回。</td></tr>
+      <tr><td><code>+delete-space</code></td><td><code>wiki +delete-space</code></td><td><span class="tag full">超过</span></td><td>预读目标、高风险确认、只接受 <code>success=true</code>；兼容 <code>+space-delete</code>。</td></tr>
+      <tr><td><code>+member-add</code></td><td><code>wiki +member-add</code></td><td><span class="tag partial">任务对齐</span></td><td>支持 1–30 个 userId 与四种角色；以写接口终态作为成功证据，不把最多 50 条的名单误作精确读回。</td></tr>
+      <tr><td><code>+member-list</code></td><td><code>wiki +member-list</code></td><td><span class="tag partial">任务对齐</span></td><td>严格成员数组、角色过滤、真实上限 50；后端无游标，不能提供诚实的 page-all。</td></tr>
+      <tr><td><code>+member-remove</code></td><td><code>wiki +member-remove</code></td><td><span class="tag partial">任务对齐</span></td><td>支持批量 userId；只接受写接口明确终态，并公开无法进行精确成员读回的边界。</td></tr>
+      <tr><td><code>+node-list</code></td><td><code>wiki +node-list</code></td><td><span class="tag full">完整对齐</span></td><td>正确跨域路由 doc/list_nodes，严格空目录、分页与自动翻页。</td></tr>
+      <tr><td><code>+node-get</code></td><td><code>wiki +node-get</code></td><td><span class="tag partial">任务对齐</span></td><td>支持 DingTalk 节点 ID/在线文档 URL 并返回文档域元数据；不接受跨平台专用的 token/type 组合。</td></tr>
+      <tr><td><code>+node-create</code></td><td><code>wiki +node-create</code></td><td><span class="tag partial">任务对齐</span></td><td>支持 adoc/axls/able/appt/adraw/amind/folder 并读回；无 origin/shortcut 等价接口。</td></tr>
+      <tr><td><code>+node-copy</code></td><td><code>wiki +node-copy</code></td><td><span class="tag full">超过</span></td><td>确认后要求新 nodeId 并读取副本；底层面向在线节点，不把 .dlink 当独立副本。</td></tr>
+      <tr><td><code>+move</code></td><td><code>wiki +move</code></td><td><span class="tag partial">任务对齐</span></td><td>同一入口支持 Wiki 内移动和“我的文档”在线节点入 Wiki，读回 workspace/folder；底层接口没有 apply 权限迁移开关。</td></tr>
+      <tr><td><code>+move-to-drive</code></td><td><code>wiki +move-to-drive</code></td><td><span class="tag full">超过</span></td><td>DWS 当前接口同步完成并读回 workspace 变化，无需暴露异步 task 轮询。</td></tr>
+      <tr><td><code>+node-delete</code></td><td><code>wiki +node-delete</code></td><td><span class="tag full">超过</span></td><td>预读并核对 workspace，高风险确认，要求删除终态。</td></tr>
+    </tbody></table></div>
+  </section>
+
+  <section class="card">
+    <h2>DWS 可挖掘的 7 个额外场景</h2>
+    <div class="two">
+      <div><h3>定位与创建链</h3><ul><li><code>+space-search</code>：严格关键词搜索。</li><li><code>+space-get</code>：空间详情与 workspaceId 证据。</li><li><code>+resolve-space</code>：唯一命中直出 ID，多命中拒绝猜测。</li><li><code>+wiki-new-doc</code>：空间名解析 → 创建 → 文档读回。</li></ul></div>
+      <div><h3>治理与巡检链</h3><ul><li><code>+member-update</code>：角色变更终态与不可精确读回声明。</li><li><code>+node-search</code>：库内关键词/扩展名搜索，严格零命中。</li><li><code>+feed-list</code>：知识库动态时间线与服务端 exclude-file 过滤。</li></ul></div>
+    </div>
+  </section>
+
+  <section class="card">
+    <h2>隐藏问题与修复</h2>
+    <table><thead><tr><th>原问题</th><th>错误风险</th><th>本次修复</th></tr></thead><tbody>
+      <tr><td>5 个旧 Wiki Shortcut 可直接执行，但只有 1 个进入公开目录。</td><td>Help、Schema、Skill 发现链与运行面漂移。</td><td><span class="tag fixed">20 项统一评审</span> 全部具备 Contract/Safety/Result 与语义目录记录。</td></tr>
+      <tr><td>列表投影找不到数组或遇到坏元素时返回空 slice。</td><td>把内部错误、字段漂移误报为“没有数据”。</td><td><span class="tag fixed">失败关闭</span> 只有响应中真实存在的 <code>[]</code> 才是合法空集合。</td></tr>
+      <tr><td>节点列表 Shortcut 调错 Wiki MCP 服务。</td><td>真实后端 <code>success=false</code>，Mock/静态检查看不出。</td><td><span class="tag fixed">跨域路由</span> 明确调用 doc/list_nodes，并纳入真实 E2E。</td></tr>
+      <tr><td>成员帮助宣称最大 200。</td><td>真实接口超过 50 直接参数错误。</td><td><span class="tag fixed">真实上限</span> Shortcut 与原子 Help 均改为 50，并在本地提前拒绝。</td></tr>
+      <tr><td>成员写操作从最多 50 条、不可分页的名单推断成员存在或缺失。</td><td>目标在截断部分时会误报写失败，或把未验证的移除报告为已读回。</td><td><span class="tag fixed">终态证据</span> 只接受写接口 <code>success=true</code>，并在结果中明确 <code>readbackAvailable=false</code>。</td></tr>
+      <tr><td>空间搜索的稳定工作流属性名与实际请求属性名不同。</td><td>直接改写已发布的 <code>query/limit</code> 会造成无版本 Schema 破坏；继续隐式转换又会让审计者误以为请求同名透传。</td><td><span class="tag fixed">显式复合适配</span> 最终 Schema 保留兼容属性并明确声明转换为 <code>keyword/pageSize</code>；回归测试同时锁定最终交付和精确请求参数。</td></tr>
+      <tr><td>知识库名称帮助宣称最大 100。</td><td>真实接口超过 32 失败。</td><td><span class="tag fixed">真实上限</span> Help 与 Shortcut 校验统一为 32。</td></tr>
+      <tr><td>复制/移动/创建只把无异常视为成功。</td><td>空确认、未知远端效果或移动未到目标仍可能被接受。</td><td><span class="tag fixed">读回证明</span> 在后端具备精确查询能力时检查 success、业务 ID、workspace/folder 等最终状态。</td></tr>
+    </tbody></table>
+  </section>
+
+  <section class="card">
+    <h2>真实数据 E2E 证据矩阵</h2>
+    <p class="small">28 项业务断言全部通过。测试使用一次性空知识库、临时在线文档与一名同组织内部测试成员；所有对象在 finally 清理。报告不保存对象 ID、成员身份、组织信息、URL、trace 或原始响应。</p>
+    <div class="toolbar"><input id="q" placeholder="筛选命令或证据"><select id="g"><option value="">全部分组</option><option>空间</option><option>成员</option><option>节点</option><option>动态</option></select></div>
+    <div class="matrix"><table id="catalog"><thead><tr><th>分组</th><th>Shortcut</th><th>实际业务断言</th><th>状态</th></tr></thead><tbody>
+      <tr><td>空间</td><td><code>+space-list</code></td><td>真实 count、hasMore、nextCursor；自动翻页返回两页结果。</td><td><span class="tag full">PASS</span></td></tr>
+      <tr><td>空间</td><td><code>+space-search</code></td><td>等待搜索索引后命中一次性 workspaceId。</td><td><span class="tag full">PASS</span></td></tr>
+      <tr><td>空间</td><td><code>+space-get</code></td><td>读回 workspaceId 与创建结果一致。</td><td><span class="tag full">PASS</span></td></tr>
+      <tr><td>空间</td><td><code>+resolve-space</code></td><td>唯一名称解析为同一 workspaceId。</td><td><span class="tag full">PASS</span></td></tr>
+      <tr><td>空间</td><td><code>+space-create</code></td><td>success=true、workspaceId 非空、详情读回一致。</td><td><span class="tag full">PASS</span></td></tr>
+      <tr><td>空间</td><td><code>+delete-space</code></td><td>目标预读、确认、success=true；兼容别名执行 finally 清理。</td><td><span class="tag full">PASS</span></td></tr>
+      <tr><td>成员</td><td><code>+member-list</code></td><td>真实 owner 条目与显式 members 数组，limit=50。</td><td><span class="tag full">PASS</span></td></tr>
+      <tr><td>成员</td><td><code>+member-add</code></td><td>命令只报告写终态；一次性小规模空间另行确认名单完整且角色为 READER。</td><td><span class="tag full">PASS</span></td></tr>
+      <tr><td>成员</td><td><code>+member-update</code></td><td>命令只报告写终态；一次性小规模空间另行确认角色变为 EDITOR。</td><td><span class="tag full">PASS</span></td></tr>
+      <tr><td>成员</td><td><code>+member-remove</code></td><td>命令只报告写终态；一次性小规模空间另行确认完整名单中不存在该 userId。</td><td><span class="tag full">PASS</span></td></tr>
+      <tr><td>节点</td><td><code>+node-list</code></td><td>空库返回真实 nodes:[]；有数据时验证游标与自动翻页。</td><td><span class="tag full">PASS</span></td></tr>
+      <tr><td>节点</td><td><code>+node-get</code></td><td>读回 nodeId 与请求一致。</td><td><span class="tag full">PASS</span></td></tr>
+      <tr><td>节点</td><td><code>+node-search</code></td><td>等待索引后按标题命中真实 nodeId。</td><td><span class="tag full">PASS</span></td></tr>
+      <tr><td>节点</td><td><code>+node-create</code></td><td>分别创建 folder/adoc，均取得 nodeId 和元数据读回。</td><td><span class="tag full">PASS</span></td></tr>
+      <tr><td>节点</td><td><code>+node-copy</code></td><td>取得不同的新 nodeId，副本元数据可读。</td><td><span class="tag full">PASS</span></td></tr>
+      <tr><td>节点</td><td><code>+move</code></td><td>读回 workspaceId 与 folderId 均等于目标；兼容 +node-move。</td><td><span class="tag full">PASS</span></td></tr>
+      <tr><td>节点</td><td><code>+move-to-drive</code></td><td>移动后读回 workspace 发生变化，再通过 +move 移回。</td><td><span class="tag full">PASS</span></td></tr>
+      <tr><td>节点</td><td><code>+node-delete</code></td><td>目标预读与 workspace 核对后收到 success=true。</td><td><span class="tag full">PASS</span></td></tr>
+      <tr><td>节点</td><td><code>+wiki-new-doc</code></td><td>按唯一空间名创建，nodeId 与文档详情读回一致。</td><td><span class="tag full">PASS</span></td></tr>
+      <tr><td>动态</td><td><code>+feed-list</code></td><td>创建/移动操作后返回真实 feeds 数组，缺字段不会被接受。</td><td><span class="tag full">PASS</span></td></tr>
+    </tbody></table></div>
+    <p class="small">可复跑入口：<code>make build</code> 后设置临时 <code>DWS_WIKI_E2E_MEMBER_ID</code>，在交互终端运行 <code>./scripts/dev/wiki-shortcut-e2e.py</code>。脚本只输出能力标签，不输出业务对象；受保护操作及最终清理均由命令逐项获取终端确认，非交互环境会在创建测试数据前拒绝运行。</p>
+  </section>
+
+  <section class="card">
+    <h2>成功判定与发布门</h2>
+    <div class="two"><div><h3>运行时证据层</h3><ol><li>传输/MCP 调用成功。</li><li>响应契约存在且类型正确。</li><li>写操作必须有 <code>success=true</code>；创建类操作还必须有业务 ID。</li><li>后端具备精确查询时必须读回；不具备时明确发布不可读回，而非从截断集合推断。</li><li>集合只有显式数组才允许为空。</li></ol></div><div><h3>交付门</h3><ol><li>20/20 语义目录与注册面精确覆盖。</li><li>Contract、Safety、Result、统一输出完整。</li><li>生成漂移、Schema、确认真值、全量 Go 测试。</li><li>独立真实数据 E2E 与 finally 清理。</li><li>diff PII/密钥/本地绝对路径扫描。</li></ol></div></div>
+  </section>
+  <p class="footer">DWS Wiki Shortcut business review · sanitized engineering artifact</p>
+</main>
+<script>
+  const q=document.querySelector('#q'),g=document.querySelector('#g'),rows=[...document.querySelectorAll('#catalog tbody tr')];
+  function filter(){const text=q.value.trim().toLowerCase(),group=g.value;rows.forEach(r=>{const okText=!text||r.textContent.toLowerCase().includes(text),okGroup=!group||r.children[0].textContent===group;r.style.display=okText&&okGroup?'':'none'})}q.addEventListener('input',filter);g.addEventListener('change',filter);
+</script></body></html>

go.mod

@@ -2,6 +2,8 @@ module github.com/DingTalk-Real-AI/dingtalk-workspace-cli
 
 go 1.25.9
 
+replace gitlab.alibaba-inc.com/aes/aem-go-sdk => ./third_party/aem-go-sdk
+
 require (
 	github.com/Microsoft/go-winio v0.6.2
 	github.com/RealAlexandreAI/json-repair v0.0.15
@@ -16,7 +18,9 @@ require (
 	github.com/muesli/termenv v0.16.0
 	github.com/open-dingtalk/dingtalk-stream-sdk-go v0.9.2-beta.1
 	github.com/spf13/cobra v1.10.2
+	github.com/yuin/goldmark v1.8.5
 	github.com/zalando/go-keyring v0.2.8
+	gitlab.alibaba-inc.com/aes/aem-go-sdk v0.3.0
 	golang.org/x/crypto v0.49.0
 	golang.org/x/sys v0.42.0
 	golang.org/x/text v0.35.0

go.sum

@@ -105,6 +105,8 @@ github.com/stretchr/testify v1.11.1 h1:7s2iGBzp5EwR7/aIZr8ao5+dra3wiQyKjjFuvgVKu
 github.com/stretchr/testify v1.11.1/go.mod h1:wZwfW3scLgRK+23gO65QZefKpKQRnfz6sD981Nm4B6U=
 github.com/xo/terminfo v0.0.0-20220910002029-abceb7e1c41e h1:JVG44RsyaB9T2KIHavMF/ppJZNG9ZpyihvCd0w101no=
 github.com/xo/terminfo v0.0.0-20220910002029-abceb7e1c41e/go.mod h1:RbqR21r5mrJuqunuUZ/Dhy/avygyECGrLceyNeo4LiM=
+github.com/yuin/goldmark v1.8.5 h1:r6N5afV5qj/5S4UTch8agZHJ8UxNCMwX7WjkkJam2NA=
+github.com/yuin/goldmark v1.8.5/go.mod h1:ip/1k0VRfGynBgxOz0yCqHrbZXhcjxyuS66Brc7iBKg=
 github.com/zalando/go-keyring v0.2.8 h1:6sD/Ucpl7jNq10rM2pgqTs0sZ9V3qMrqfIIy5YPccHs=
 github.com/zalando/go-keyring v0.2.8/go.mod h1:tsMo+VpRq5NGyKfxoBVjCuMrG47yj8cmakZDO5QGii0=
 go.yaml.in/yaml/v3 v3.0.4/go.mod h1:DhzuOOF2ATzADvBadXxruRBLzYTpT36CKvDb3+aBEFg=

internal/app/event_command.go

@@ -403,7 +403,7 @@ SIGTERM、关 stdin，或先用 dws event stop <subscribe_id> --dry-run 预览
 			Selection: contract.SelectionSpec{
 				AgentSummary: "消费 OA、群生命周期或需要底层控制的个人事件流；Agent 通常使用 --flatten 输出 NDJSON",
 				UseWhen: []string{
-					"需要监听六个公开 OA 审批任务/实例 EventKey 中的一个或多个事件",
+					"需要监听七个公开 OA 审批任务/实例 EventKey 中的一个或多个事件",
 					"需要监听指定群的标题变更、成员进退群或群解散事件",
 					"用户显式给出原始 EventKey、Filter DSL、subscribe_id，要求原始 transport envelope，或需要普通 IM facade 不提供的高级多事件控制",
 				},
@@ -1220,7 +1220,8 @@ func newEventStopCommandWithFlags(globalFlags ...*GlobalFlags) *cobra.Command {
 			editionName := editionNameOrDefault()
 			clientIDHash := dwsevent.ClientIDHash(clientID)
 			workDir := eventWorkDir(configDir, editionName, dwsevent.SourceKindAppStream, clientIDHash)
-			if err := eventStopBus(busctl.StopConfig{WorkDir: workDir}); err != nil {
+			ipcEndpoint := defaultIPCEndpoint(workDir, editionName, dwsevent.SourceKindAppStream, clientIDHash)
+			if err := eventStopBus(busctl.StopConfig{WorkDir: workDir, IPCEndpoint: ipcEndpoint}); err != nil {
 				if errors.Is(err, busctl.ErrNotRunning) {
 					fmt.Fprintln(c.OutOrStdout(), "bus is not running")
 					return nil

internal/app/event_personal_attempts_test.go

@@ -152,8 +152,8 @@ func TestCrossPlatformCoveragePersonalSubscriptionProtectionCoversAllPublicEvent
 		}
 	}
 
-	if publicCount != 22 {
-		t.Fatalf("public personal events = %d, want 22 (16 IM + 6 OA)", publicCount)
+	if publicCount != 23 {
+		t.Fatalf("public personal events = %d, want 23 (16 IM + 7 OA)", publicCount)
 	}
 	for _, ruleType := range []string{"at", "all", "singleChat", "sender", "group"} {
 		if !ruleTypes[ruleType] {

internal/app/event_personal_command.go

@@ -1220,7 +1220,7 @@ func runPersonalEventStop(c *cobra.Command, opts personalStopOptions) error {
 	}
 
 	busState := "personal bus stopped"
-	if err := personalStopBus(busctl.StopConfig{WorkDir: workDir}); err != nil {
+	if err := personalStopBus(busctl.StopConfig{WorkDir: workDir, IPCEndpoint: ipcEndpoint}); err != nil {
 		if errors.Is(err, busctl.ErrNotRunning) {
 			busState = "personal bus is not running"
 		} else {

internal/app/event_personal_oa_test.go

@@ -61,6 +61,13 @@ func TestPersonalOAEventListAndSchemaCommands(t *testing.T) {
 				"process_code", "title", "status", "create_time", "event_time",
 			},
 		},
+		{
+			eventKey: personal.EventOAApprovalInstanceCC,
+			properties: []string{
+				"type", "event_id", "timestamp", "subscribe_id", "process_instance_id",
+				"process_code", "title", "status", "create_time", "event_time",
+			},
+		},
 		{
 			eventKey: personal.EventOAApprovalInstanceTerminated,
 			properties: []string{
@@ -161,6 +168,7 @@ func TestPersonalOAEventConsumeDryRunAndValidation(t *testing.T) {
 		personal.EventOAApprovalTaskFinished,
 		personal.EventOAApprovalTaskRedirected,
 		personal.EventOAApprovalInstanceStarted,
+		personal.EventOAApprovalInstanceCC,
 		personal.EventOAApprovalInstanceTerminated,
 		personal.EventOAApprovalInstanceFinished,
 	}
@@ -414,6 +422,7 @@ func TestPersonalOAMultiConsumeCreatesIndependentAllSubscriptionsOnSharedBus(t *
 		personal.EventOAApprovalTaskFinished,
 		personal.EventOAApprovalTaskRedirected,
 		personal.EventOAApprovalInstanceStarted,
+		personal.EventOAApprovalInstanceCC,
 		personal.EventOAApprovalInstanceTerminated,
 		personal.EventOAApprovalInstanceFinished,
 	}

internal/app/flag_error_test.go

@@ -82,6 +82,46 @@ func TestFlagErrorWithSuggestions_unknownFlagHintAndFlags(t *testing.T) {
 	}
 }
 
+func TestFlagErrorWithSuggestionsChatFromExplainsBothMeanings(t *testing.T) {
+	t.Parallel()
+	root := &cobra.Command{Use: "dws"}
+	chat := &cobra.Command{Use: "chat"}
+	search := &cobra.Command{Use: "+search-msg", Run: func(*cobra.Command, []string) {}}
+	search.Flags().String("sender", "", "sender target")
+	search.Flags().String("start", "", "start time")
+	root.AddCommand(chat)
+	chat.AddCommand(search)
+
+	orig := fmt.Errorf("unknown flag: --from")
+	err := flagErrorWithSuggestions(search, orig)
+	var ae *apperrors.Error
+	if !stderrors.As(err, &ae) {
+		t.Fatalf("want *apperrors.Error, got %T", err)
+	}
+	if ae.Reason != "ambiguous_flag" || !strings.Contains(ae.Hint, "--sender") || !strings.Contains(ae.Hint, "--start") {
+		t.Fatalf("structured error = reason %q hint %q", ae.Reason, ae.Hint)
+	}
+	if !strings.HasSuffix(ae.Message, "See 'dws chat +search-msg --help' for usage.") {
+		t.Fatalf("Message = %q", ae.Message)
+	}
+
+	for _, flag := range []string{"from-file", "from-user"} {
+		t.Run(flag, func(t *testing.T) {
+			err := flagErrorWithSuggestions(search, fmt.Errorf("unknown flag: --%s", flag))
+			var structured *apperrors.Error
+			if stderrors.As(err, &structured) && structured.Reason == "ambiguous_flag" {
+				t.Fatalf("--%s incorrectly used --from ambiguity handling: %#v", flag, structured)
+			}
+			if strings.Contains(err.Error(), "--from 在消息查询中含义不明确") {
+				t.Fatalf("--%s incorrectly received --from ambiguity hint: %v", flag, err)
+			}
+			if !strings.Contains(err.Error(), "unknown flag: --"+flag) {
+				t.Fatalf("error = %q, want original flag --%s", err, flag)
+			}
+		})
+	}
+}
+
 // TestFlagErrorWithSuggestions_fallbackTailHint 验证 fallback 路径（非 unknown flag 类错误，
 // 如 missing required flag / ambiguous shorthand）也带尾部 See '<cmd> --help' for usage.
 // 这是 wukong / docker / kubectl 的通用 UX——任何 flag 解析错误都给用户一条 help 入口。

internal/app/framework_coverage_test.go

@@ -268,7 +268,7 @@ type frameworkFailWriter struct{}
 
 func (frameworkFailWriter) Write([]byte) (int, error) { return 0, errors.New("write failed") }
 
-func TestFrameworkExecutePanicBeforeEmissionUsesUnifiedFailure(t *testing.T) {
+func TestCrossPlatformCoverageFrameworkExecutePanicBeforeEmissionUsesUnifiedFailure(t *testing.T) {
 	for _, failWriter := range []bool{false, true} {
 		t.Run(map[bool]string{false: "emits", true: "fallback"}[failWriter], func(t *testing.T) {
 			testseam.Protect(t, &os.Args)
@@ -470,6 +470,118 @@ func TestCrossPlatformCoverageFrameworkExecuteRareOutcomeBranches(t *testing.T)
 	})
 }
 
+func TestCrossPlatformCoverageExecuteDeterministicInterruptionBranches(t *testing.T) {
+	install := func(t *testing.T, state *processSignalState, stdout, stderr io.Writer) {
+		t.Helper()
+		testseam.Protect(t, &os.Args)
+		os.Args = []string{"dws"}
+		testseam.Swap(t, &rootNormalizeProcessProfileArgs, func() func() { return func() {} })
+		testseam.Swap(t, &rootStopAllStdioClients, func() {})
+		testseam.Swap(t, &rootInstallProcessSignalContext, func(ctx context.Context, _ *output.ResultStore) (context.Context, *processSignalState, func()) {
+			return ctx, state, func() {}
+		})
+		testseam.Swap(t, &rootNewRootCommandWithEngine, func(ctx context.Context, _ *pipeline.Engine) *cobra.Command {
+			cmd := &cobra.Command{Use: "dws", SilenceErrors: true, SilenceUsage: true}
+			output.SetCommandRollout(cmd, output.RolloutUnifiedActive)
+			cmd.SetContext(ctx)
+			cmd.SetOut(stdout)
+			cmd.SetErr(stderr)
+			return cmd
+		})
+	}
+	interrupted := func(primaryCompleted bool) *processSignalState {
+		return &processSignalState{
+			interruption:             &processInterruption{signal: os.Interrupt},
+			primaryCompletedAtSignal: primaryCompleted,
+		}
+	}
+
+	t.Run("preparse interruption emits unified failure", func(t *testing.T) {
+		var stdout bytes.Buffer
+		install(t, interrupted(false), &stdout, io.Discard)
+		testseam.Swap(t, &rootRunPreParse, func(*cobra.Command, *pipeline.Engine) error { return errors.New("preparse failed") })
+		testseam.Swap(t, &rootExecuteCommand, func(*cobra.Command) (*cobra.Command, error) {
+			t.Fatal("preparse failure reached command execution")
+			return nil, nil
+		})
+		if code, _, summary := ExecuteWithTelemetry(); code != 130 || summary == "" || !strings.Contains(stdout.String(), `"outcome": "failure"`) {
+			t.Fatalf("preparse interruption = code %d summary %q stdout %q", code, summary, stdout.String())
+		}
+	})
+
+	t.Run("interruption before emission becomes primary error", func(t *testing.T) {
+		var stdout bytes.Buffer
+		install(t, interrupted(false), &stdout, io.Discard)
+		testseam.Swap(t, &rootRunPreParse, func(*cobra.Command, *pipeline.Engine) error { return nil })
+		testseam.Swap(t, &rootExecuteCommand, func(cmd *cobra.Command) (*cobra.Command, error) { return cmd, nil })
+		if code, _, summary := ExecuteWithTelemetry(); code != 130 || summary == "" || !strings.Contains(stdout.String(), `"outcome": "failure"`) {
+			t.Fatalf("pre-emission interruption = code %d summary %q stdout %q", code, summary, stdout.String())
+		}
+	})
+
+	t.Run("late hook error preserves emitted result", func(t *testing.T) {
+		install(t, interrupted(true), io.Discard, io.Discard)
+		testseam.Swap(t, &rootRunPreParse, func(*cobra.Command, *pipeline.Engine) error { return nil })
+		testseam.Swap(t, &rootExecuteCommand, func(cmd *cobra.Command) (*cobra.Command, error) {
+			if err := output.StoreResult(cmd.Context(), output.Success(map[string]any{"ok": true})); err != nil {
+				t.Fatal(err)
+			}
+			if _, _, err := output.EmitStoredResult(cmd); err != nil {
+				t.Fatal(err)
+			}
+			return cmd, errors.New("late hook failed")
+		})
+		if code, _, summary := ExecuteWithTelemetry(); code != 0 || summary != "late hook failed" {
+			t.Fatalf("late hook result = code %d summary %q", code, summary)
+		}
+	})
+
+	t.Run("interruption after emission preserves emitted result", func(t *testing.T) {
+		install(t, interrupted(false), io.Discard, io.Discard)
+		testseam.Swap(t, &rootRunPreParse, func(*cobra.Command, *pipeline.Engine) error { return nil })
+		testseam.Swap(t, &rootExecuteCommand, func(cmd *cobra.Command) (*cobra.Command, error) {
+			if err := output.StoreResult(cmd.Context(), output.Success(map[string]any{"ok": true})); err != nil {
+				t.Fatal(err)
+			}
+			if _, _, err := output.EmitStoredResult(cmd); err != nil {
+				t.Fatal(err)
+			}
+			return cmd, nil
+		})
+		if code, _, summary := ExecuteWithTelemetry(); code != 0 || summary == "" {
+			t.Fatalf("post-emission interruption = code %d summary %q", code, summary)
+		}
+	})
+
+	t.Run("publication failure replaces unobservable result", func(t *testing.T) {
+		var original bytes.Buffer
+		install(t, interrupted(false), io.Discard, io.Discard)
+		testseam.Swap(t, &rootRunPreParse, func(*cobra.Command, *pipeline.Engine) error { return nil })
+		testseam.Swap(t, &rootExecuteCommand, func(cmd *cobra.Command) (*cobra.Command, error) {
+			if err := output.StoreResult(cmd.Context(), output.Success(map[string]any{"ok": true})); err != nil {
+				t.Fatal(err)
+			}
+			if _, _, err := output.EmitStoredResult(cmd); err != nil {
+				t.Fatal(err)
+			}
+			file, err := os.CreateTemp(t.TempDir(), "finished-output-*")
+			if err != nil {
+				t.Fatal(err)
+			}
+			t.Cleanup(func() { _ = file.Close() })
+			cmd.SetContext(context.WithValue(cmd.Context(), outputFileContextKey{}, &outputSinkState{file: file, original: &original, finished: true}))
+			publicationErr := newOutputPublicationError("publish", errors.New("rename failed"))
+			if _, handled, emitErr := emitOutputPublicationFailure(cmd, publicationErr); !handled || emitErr != nil {
+				t.Fatalf("precondition publication failure = handled %v error %v unified %v state %v", handled, emitErr, output.UsesUnifiedResult(cmd), outputSinkForCommand(cmd) != nil)
+			}
+			return cmd, publicationErr
+		})
+		if code, _, summary := ExecuteWithTelemetry(); code != 5 || summary == "" {
+			t.Fatalf("publication failure = code %d summary %q output %q", code, summary, original.String())
+		}
+	})
+}
+
 type frameworkPanicWriter struct{}
 
 func (frameworkPanicWriter) Write([]byte) (int, error) { panic("writer panic") }

internal/app/param_alias_e2e_test.go

@@ -51,7 +51,40 @@ func (c *paramAliasCaptureCaller) CallTool(_ context.Context, server, tool strin
 func (c *paramAliasCaptureCaller) paramAliasResponseForTool(tool string) string {
 	switch tool {
 	case "list_calendar_events":
-		return `{"result":{"events":[]}}`
+		return `{"success":true,"result":{"events":[],"hasMore":false,"nextCursor":""}}`
+	case "get_calendar_detail":
+		return c.paramAliasCalendarDetailResponse()
+	case "get_calendar_participants":
+		return `{"success":true,"result":{"participants":[{"userId":"fixture-user","displayName":"Fixture User"},{"userId":"user-2","displayName":"User Two"}]}}`
+	case "search_calendar":
+		return `{"success":true,"result":{"calendars":[]}}`
+	case "search_rooms":
+		return `{"success":true,"result":{"rooms":[]}}`
+	case "query_available_meeting_room":
+		return `{"success":true,"result":{"rooms":[],"hasMore":false}}`
+	case "list_meeting_room_groups":
+		return `{"success":true,"result":{"groups":[]}}`
+	case "query_busy_status":
+		return `{"success":true,"result":[]}`
+	case "list_suggested_event_times":
+		return `{"success":true,"result":{"recommendEventTimes":[]}}`
+	case "create_calendar_event":
+		return `{"success":true,"result":{"eventId":"event-1"}}`
+	case "update_calendar_event", "delete_calendar_event", "add_calendar_participant", "remove_calendar_participant":
+		return `{"success":true}`
+	case "respond":
+		status := "accepted"
+		if call := c.lastParamAliasCall(); call != nil {
+			if value, ok := call.args["responseStatus"].(string); ok && value != "" {
+				status = value
+			}
+		}
+		encoded, _ := json.Marshal(map[string]any{"success": true, "result": map[string]any{"responseStatus": status}})
+		return string(encoded)
+	case "get_current_user_profile":
+		return `{"success":true,"result":{"userId":"user-1","name":"Fixture Current User"}}`
+	case "query_records":
+		return `{"success":true,"status":"success","error":{},"data":{}}`
 	case "search_mail_users":
 		return `{"users":[{"name":"Fixture User","email":"fixture@example.com","id":"fixture-user"}]}`
 	case "search_dept_by_keyword":
@@ -63,9 +96,11 @@ func (c *paramAliasCaptureCaller) paramAliasResponseForTool(tool string) string
 	case "list_doc_versions":
 		return `{"result":{"items":[{"version":3}]}}`
 	case "revert_doc_version":
-		return `{"version":3}`
+		return `{"revertedToVersion":3}`
 	case "search_doc_templates":
 		return `{"result":[{"templateId":"fixture-template-id"}]}`
+	case "list_workflows":
+		return `{"workflows":[]}`
 	case "create_document":
 		return `{"nodeId":"fixture-node"}`
 	case "list_files":
@@ -123,6 +158,41 @@ func (c *paramAliasCaptureCaller) paramAliasResponseForTool(tool string) string
 	}
 }
 
+func (c *paramAliasCaptureCaller) lastParamAliasCall() *paramAliasToolCall {
+	if len(c.calls) == 0 {
+		return nil
+	}
+	return &c.calls[len(c.calls)-1]
+}
+
+func (c *paramAliasCaptureCaller) paramAliasCalendarDetailResponse() string {
+	event := map[string]any{
+		"eventId":       "event-1",
+		"summary":       "Fixture Meeting",
+		"description":   "fixture description",
+		"startDateTime": "2026-03-10T09:00:00+08:00",
+		"endDateTime":   "2026-03-10T10:00:00+08:00",
+	}
+	for _, call := range c.calls {
+		switch call.tool {
+		case "create_calendar_event", "update_calendar_event":
+			for _, key := range []string{"eventId", "summary", "description", "startDateTime", "endDateTime", "timeZone", "location", "freeBusy"} {
+				if value, ok := call.args[key]; ok {
+					event[key] = value
+				}
+			}
+		case "respond":
+			if value, ok := call.args["responseStatus"]; ok {
+				event["responseStatus"] = value
+			}
+		case "delete_calendar_event":
+			event["status"] = "cancelled"
+		}
+	}
+	encoded, _ := json.Marshal(map[string]any{"success": true, "result": event})
+	return string(encoded)
+}
+
 func (*paramAliasCaptureCaller) Format() string { return "json" }
 func (*paramAliasCaptureCaller) DryRun() bool   { return false }
 func (*paramAliasCaptureCaller) Fields() string { return "" }

internal/app/param_alias_payload_equivalence_test.go

@@ -5,6 +5,8 @@ package app
 
 import (
 	"errors"
+	"os"
+	"os/exec"
 	"reflect"
 	"strings"
 	"testing"
@@ -17,6 +19,8 @@ import (
 const (
 	appFixtureCurrentDOpenID  = "DAAAAAAAAAAAiE"
 	appFixtureCurrentDOpenID2 = "DAQEBAQEBAQEiE"
+
+	paramAliasCalendarPayloadChildEnv = "DWS_TEST_CALENDAR_PARAM_ALIAS_PAYLOAD_CHILD"
 )
 
 // paramAliasCompleteCommands is deliberately keyed by the exact reviewed
@@ -27,15 +31,56 @@ const (
 // its spelling while holding every other input constant.
 var paramAliasCompleteCommands = map[string][]string{
 	"aitable +base-search":                     {"aitable", "+base-search", "--query", "fixture"},
+	"aitable +export-data":                     {"aitable", "+export-data", "--base-id", "base-1", "--scope", "all", "--format", "excel"},
 	"aitable +field-get":                       {"aitable", "+field-get", "--base-id", "base-1", "--table-id", "table-1"},
+	"aitable +find-record":                     {"aitable", "+find-record", "--base", "base-1", "--table", "table-1", "--query", "fixture"},
 	"aitable +list-tables":                     {"aitable", "+list-tables", "--base", "base-1"},
 	"aitable +record-query":                    {"aitable", "+record-query", "--base-id", "base-1", "--table-id", "table-1", "--query", "fixture"},
+	"aitable +record-share-links":              {"aitable", "+record-share-links", "--base", "base-1", "--table", "table-1", "--record-ids", "record-1"},
 	"aitable +record-share-url":                {"aitable", "+record-share-url", "--base-id", "base-1", "--table-id", "table-1", "--record-ids", "record-1"},
 	"aitable +table-get":                       {"aitable", "+table-get", "--base-id", "base-1"},
+	"aitable +workflow-list":                   {"aitable", "+workflow-list", "--base-id", "base-1", "--limit", "7"},
+	"aitable attachment upload":                {"aitable", "attachment", "upload", "--base-id", "base-1", "--file-name", "fixture.txt", "--size", "7"},
+	"aitable base list":                        {"aitable", "base", "list", "--cursor", "cursor-1", "--limit", "7"},
+	"aitable base update":                      {"aitable", "base", "update", "--base-id", "base-1", "--name", "Fixture Base", "--desc", "fixture description"},
+	"aitable field search-options":             {"aitable", "field", "search-options", "--base-id", "base-1", "--table-id", "table-1", "--field-id", "field-1", "--keyword", "fixture", "--limit", "7"},
 	"aitable record query":                     {"aitable", "record", "query", "--base-id", "base-1", "--table-id", "table-1", "--limit", "7"},
+	"aitable workflow get":                     {"aitable", "workflow", "get", "--base-id", "base-1", "--workflow-id", "workflow-1"},
+	"aitable workflow history":                 {"aitable", "workflow", "history", "--base-id", "base-1", "--workflow-id", "workflow-1", "--after-time", "1000", "--before-time", "2000", "--page", "2", "--size", "25"},
+	"aitable workflow run":                     {"aitable", "workflow", "run", "--base-id", "base-1", "--workflow-id", "workflow-1", "--table-id", "table-1", "--record-ids", "record-1", "--yes"},
 	"attendance check result":                  {"attendance", "check", "result", "--users", "user-1,user-2", "--start", "2026-03-01", "--end", "2026-03-02"},
 	"attendance +check-result":                 {"attendance", "+check-result", "--users", "user-1,user-2", "--start", "2026-03-01", "--end", "2026-03-02"},
+	"calendar +agenda":                         {"calendar", "+agenda", "--start", "2026-03-10T09:00:00+08:00", "--end", "2026-03-10T18:00:00+08:00", "--calendar-id", "primary", "--cursor", "cursor-1", "--limit", "7"},
+	"calendar +attendee-list":                  {"calendar", "+attendee-list", "--event", "event-1", "--calendar-id", "primary"},
+	"calendar +book":                           {"calendar", "+book", "--title", "Fixture Meeting", "--start", "2026-03-10T09:00:00+08:00", "--end", "2026-03-10T10:00:00+08:00", "--with", "Fixture User", "--yes"},
+	"calendar +book-search":                    {"calendar", "+book-search", "--query", "fixture"},
+	"calendar +cancel-event":                   {"calendar", "+cancel-event", "--event", "event-1", "--yes"},
+	"calendar +conflicts":                      {"calendar", "+conflicts", "--in-days", "1"},
+	"calendar +create":                         {"calendar", "+create", "--title", "Fixture Meeting", "--start", "2026-03-10T09:00:00+08:00", "--end", "2026-03-10T10:00:00+08:00", "--desc", "fixture description", "--attendees", "user-1,user-2", "--rooms", "room-1,room-2", "--calendar-id", "primary", "--yes"},
+	"calendar +free":                           {"calendar", "+free", "--who", "Fixture User", "--start", "2026-03-10T09:00:00+08:00", "--end", "2026-03-10T18:00:00+08:00"},
+	"calendar +free-slots":                     {"calendar", "+free-slots", "--from", "9", "--to", "18", "--in-days", "1"},
+	"calendar +freebusy":                       {"calendar", "+freebusy", "--users", "user-1,user-2", "--rooms", "room-1,room-2", "--start", "2026-03-10T09:00:00+08:00", "--end", "2026-03-10T18:00:00+08:00"},
+	"calendar +get":                            {"calendar", "+get", "--event", "event-1", "--calendar-id", "primary"},
+	"calendar +invite":                         {"calendar", "+invite", "--event", "event-1", "--with", "Fixture User", "--yes"},
+	"calendar +my-free":                        {"calendar", "+my-free", "--start", "2026-03-10T09:00:00+08:00", "--end", "2026-03-10T18:00:00+08:00"},
+	"calendar +reschedule":                     {"calendar", "+reschedule", "--event", "event-1", "--start", "2026-03-10T10:00:00+08:00", "--end", "2026-03-10T11:00:00+08:00", "--yes"},
+	"calendar +room-find":                      {"calendar", "+room-find", "--start", "2026-03-10T09:00:00+08:00", "--end", "2026-03-10T10:00:00+08:00", "--room-name", "Fixture Room", "--group-id", "group-1", "--page", "1", "--limit", "7"},
+	"calendar +room-groups":                    {"calendar", "+room-groups", "--page", "1", "--limit", "7"},
+	"calendar +room-search":                    {"calendar", "+room-search", "--room-name", "Fixture Room"},
+	"calendar +rsvp":                           {"calendar", "+rsvp", "--event", "event-1", "--status", "accept", "--calendar-id", "primary", "--yes"},
+	"calendar +search-event":                   {"calendar", "+search-event", "--query", "fixture", "--start", "2026-03-10T09:00:00+08:00", "--end", "2026-03-10T18:00:00+08:00", "--calendar-id", "primary", "--cursor", "cursor-1", "--limit", "7"},
+	"calendar +suggest-time":                   {"calendar", "+suggest-time", "--with", "Fixture User", "--duration", "30", "--start", "2026-03-10T09:00:00+08:00", "--end", "2026-03-10T18:00:00+08:00"},
+	"calendar +suggestion":                     {"calendar", "+suggestion", "--users", "user-1,user-2", "--duration", "30", "--start", "2026-03-10T09:00:00+08:00", "--end", "2026-03-10T18:00:00+08:00", "--timezone", "Asia/Shanghai"},
+	"calendar +update":                         {"calendar", "+update", "--event", "event-1", "--title", "Fixture Updated Meeting", "--desc", "fixture updated description", "--start", "2026-03-10T10:00:00+08:00", "--end", "2026-03-10T11:00:00+08:00", "--add-attendees", "user-2", "--remove-attendees", "user-1", "--yes"},
+	"calendar busy search":                     {"calendar", "busy", "search", "--users", "user-1,user-2", "--rooms", "room-1,room-2", "--start", "2026-03-10T09:00:00+08:00", "--end", "2026-03-10T18:00:00+08:00"},
+	"calendar event create":                    {"calendar", "event", "create", "--title", "Fixture Meeting", "--start", "2026-03-10T09:00:00+08:00", "--end", "2026-03-10T10:00:00+08:00", "--remind-minutes", "15", "--timezone", "Asia/Shanghai", "--rooms", "room-1,room-2"},
 	"calendar event list":                      {"calendar", "event", "list", "--start", "2026-03-10T14:00:00+08:00", "--end", "2026-03-10T18:00:00+08:00", "--calendar-id", "primary", "--cursor", "cursor-1", "--limit", "7"},
+	"calendar event respond":                   {"calendar", "event", "respond", "--id", "event-1", "--status", "accepted"},
+	"calendar event suggest":                   {"calendar", "event", "suggest", "--users", "user-1,user-2", "--duration", "30", "--start", "2026-03-10T09:00:00+08:00", "--end", "2026-03-10T18:00:00+08:00", "--timezone", "Asia/Shanghai"},
+	"calendar event update":                    {"calendar", "event", "update", "--id", "event-1", "--timezone", "Asia/Shanghai"},
+	"calendar room add":                        {"calendar", "room", "add", "--event", "event-1", "--rooms", "room-1,room-2"},
+	"calendar room delete":                     {"calendar", "room", "delete", "--event", "event-1", "--rooms", "room-1,room-2"},
+	"calendar room search":                     {"calendar", "room", "search", "--room-name", "Fixture Room", "--group-id", "group-1", "--start", "2027-03-10T09:00:00+08:00", "--end", "2027-03-10T10:00:00+08:00", "--page", "1", "--limit", "7"},
 	"chat +chat-messages":                      {"chat", "+chat-messages", "--group", "fixture-conversation"},
 	"chat +chat-add-bot":                       {"chat", "+chat-add-bot", "--id", "fixture-conversation", "--robot-code", "robot-1", "--yes"},
 	"chat +chat-audit-join":                    {"chat", "+chat-audit-join", "--group", "fixture-conversation", "--record-id", "7", "--applicant", "user-1", "--inviter", "user-2", "--status", "AuditApprove", "--yes"},
@@ -62,12 +107,12 @@ var paramAliasCompleteCommands = map[string][]string{
 	"chat +messages-list":                      {"chat", "+messages-list", "--group", "fixture-conversation", "--time", "2026-03-10 00:00:00", "--limit", "7"},
 	"chat +messages-list-direct":               {"chat", "+messages-list-direct", "--user", "user-1", "--time", "2026-03-10 00:00:00", "--limit", "7"},
 	"chat +messages-list-unread-conversations": {"chat", "+messages-list-unread-conversations", "--count", "7", "--exclude-muted"},
-	"chat +messages-reply":                     {"chat", "+messages-reply", "--conversation-id", "fixture-conversation", "--ref-msg-id", "message-1", "--ref-sender", appFixtureCurrentDOpenID, "--text", "hello fixture", "--yes"},
+	"chat +messages-reply":                     {"chat", "+messages-reply", "--group", "fixture-conversation", "--ref-msg-id", "message-1", "--ref-sender", appFixtureCurrentDOpenID, "--content", "hello fixture", "--yes"},
 	"chat +messages-resource-download":         {"chat", "+messages-resource-download", "--resource-id", "resource-1", "--message-id", "message-1", "--open-conversation-id", "fixture-conversation", "--output", "downloads/fixture.bin"},
 	"chat +messages-set-pin":                   {"chat", "+messages-set-pin", "--open-conversation-id", "fixture-conversation", "--msg-id", "message-1", "--yes"},
-	"chat +messages-send-by-webhook":           {"chat", "+messages-send-by-webhook", "--token", "fixture-token", "--title", "Fixture Alert", "--text", "fixture", "--at-users", "user-1,user-2", "--yes"},
+	"chat +messages-send-by-webhook":           {"chat", "+messages-send-by-webhook", "--token", "fixture-token", "--title", "Fixture Alert", "--content", "fixture", "--at-users", "user-1,user-2", "--yes"},
 	"chat +search-msg":                         {"chat", "+search-msg", "--group", "fixture-conversation", "--query", "fixture", "--start", "2026-03-10T00:00:00+08:00", "--end", "2026-03-11T00:00:00+08:00", "--no-enrich"},
-	"chat +send-to-group":                      {"chat", "+send-to-group", "--group", "Fixture Group", "--text", "hello fixture", "--yes"},
+	"chat +send-to-group":                      {"chat", "+send-to-group", "--group", "Fixture Group", "--content", "hello fixture", "--yes"},
 	"chat +unread-chats":                       {"chat", "+unread-chats", "--count", "7", "--exclude-muted"},
 	"chat bot find":                            {"chat", "bot", "find", "--query", "fixture", "--limit", "7"},
 	"chat bot search":                          {"chat", "bot", "search", "--name", "Fixture Bot", "--page", "2", "--size", "7"},
@@ -93,11 +138,11 @@ var paramAliasCompleteCommands = map[string][]string{
 	"chat message list-by-ids":                 {"chat", "message", "list-by-ids", "--msg-ids", "message-1,message-2"},
 	"chat message list-unread-conversations":   {"chat", "message", "list-unread-conversations", "--count", "7", "--exclude-muted"},
 	"chat message recall":                      {"chat", "message", "recall", "--conversation-id", "fixture-conversation", "--msg-id", "message-1", "--yes"},
-	"chat message reply":                       {"chat", "message", "reply", "--conversation-id", "fixture-conversation", "--ref-msg-id", "message-1", "--ref-sender", appFixtureCurrentDOpenID, "--text", "hello fixture", "--yes"},
+	"chat message reply":                       {"chat", "message", "reply", "--group", "fixture-conversation", "--ref-msg-id", "message-1", "--ref-sender", appFixtureCurrentDOpenID, "--content", "hello fixture", "--yes"},
 	"chat message search-advanced":             {"chat", "message", "search-advanced", "--conversation-ids", "fixture-conversation", "--query", "fixture"},
-	"chat message send":                        {"chat", "message", "send", "--user", appFixtureCurrentDOpenID, "--text", "hello fixture", "--idempotency-key", "param-alias-equivalence", "--yes"},
+	"chat message send":                        {"chat", "message", "send", "--user", appFixtureCurrentDOpenID, "--content", "hello fixture", "--idempotency-key", "param-alias-equivalence", "--yes"},
 	"chat message send-by-bot":                 {"chat", "message", "send-by-bot", "--robot-code", "robot-1", "--group", "fixture-conversation", "--title", "Fixture Alert", "--text", "@user-1 @user-2 fixture", "--at-user-ids", "user-1,user-2", "--yes"},
-	"chat message send-by-webhook":             {"chat", "message", "send-by-webhook", "--token", "fixture-token", "--title", "Fixture Alert", "--text", "fixture", "--at-users", "user-1,user-2", "--yes"},
+	"chat message send-by-webhook":             {"chat", "message", "send-by-webhook", "--token", "fixture-token", "--title", "Fixture Alert", "--content", "fixture", "--at-users", "user-1,user-2", "--yes"},
 	"contact +dept-members":                    {"contact", "+dept-members", "--dept", "Fixture Dept"},
 	"contact +list-sub-depts":                  {"contact", "+list-sub-depts", "--dept", "1"},
 	"contact +resolve-dept":                    {"contact", "+resolve-dept", "--name", "Fixture Dept"},
@@ -116,7 +161,7 @@ var paramAliasCompleteCommands = map[string][]string{
 	"doc +copy":                                {"doc", "+copy", "--node", "node-1", "--workspace", "workspace-1", "--yes"},
 	"doc +create":                              {"doc", "+create", "--name", "Fixture Document", "--content", "fixture body", "--doc-format", "markdown"},
 	"doc +create-from-template":                {"doc", "+create-from-template", "--query", "fixture template", "--name", "Fixture From Template", "--folder", "folder-1", "--workspace", "workspace-1"},
-	"doc +doc-append":                          {"doc", "+doc-append", "--doc", "node-1", "--text", "fixture appendix", "--yes"},
+	"doc +doc-append":                          {"doc", "+doc-append", "--doc", "node-1", "--content", "fixture appendix", "--yes"},
 	"doc +export-submit":                       {"doc", "+export-submit", "--node", "node-1", "--export-format", "docx"},
 	"doc +fetch":                               {"doc", "+fetch", "--node", "node-1", "--scope", "section", "--start-block-id", "block-1"},
 	"doc +find-doc":                            {"doc", "+find-doc", "--query", "fixture", "--limit", "7"},
@@ -132,8 +177,8 @@ var paramAliasCompleteCommands = map[string][]string{
 	"doc +version-save":                        {"doc", "+version-save", "--node", "node-1", "--yes"},
 	"doc +update":                              {"doc", "+update", "--node", "node-1", "--command", "overwrite", "--content", `["root",{}]`, "--doc-format", "jsonml", "--expected-revision", "1", "--yes"},
 	"doc +export":                              {"doc", "+export", "--node", "node-1", "--export-format", "docx", "--output", "exports/fixture.docx"},
-	"doc block insert":                         {"doc", "block", "insert", "--node", "node-1", "--text", "fixture paragraph", "--yes"},
-	"doc block update":                         {"doc", "block", "update", "--node", "node-1", "--block-id", "block-1", "--text", "fixture paragraph", "--yes"},
+	"doc block insert":                         {"doc", "block", "insert", "--node", "node-1", "--content", "fixture paragraph", "--yes"},
+	"doc block update":                         {"doc", "block", "update", "--node", "node-1", "--block-id", "block-1", "--content", "fixture paragraph", "--yes"},
 	"doc comment create":                       {"doc", "comment", "create", "--node", "node-1", "--content", "fixture comment", "--yes"},
 	"doc comment create-inline":                {"doc", "comment", "create-inline", "--node", "node-1", "--block-id", "block-1", "--start", "0", "--end", "7", "--content", "fixture comment", "--yes"},
 	"doc comment delete":                       {"doc", "comment", "delete", "--node", "node-1", "--comment-key", "comment-1", "--yes"},
@@ -196,7 +241,7 @@ var paramAliasCompleteCommandVariants = map[string]map[string][]string{
 		"workspace": {"doc", "+copy", "--node", "node-1", "--workspace", "workspace-1", "--yes"},
 	},
 	"doc block insert": {
-		"parent-block": {"doc", "block", "insert", "--node", "node-1", "--parent-block", "parent-block-1", "--index", "0", "--text", "fixture paragraph", "--yes"},
+		"parent-block": {"doc", "block", "insert", "--node", "node-1", "--parent-block", "parent-block-1", "--index", "0", "--content", "fixture paragraph", "--yes"},
 	},
 	"doc +inspect": {
 		"include-permissions": {"doc", "+inspect", "--node", "node-1", "--include-permissions"},
@@ -219,8 +264,8 @@ var paramAliasCompleteCommandVariants = map[string]map[string][]string{
 		"sender-open-dingtalk-id": {"chat", "message", "list-by-sender", "--sender-open-dingtalk-id", appFixtureCurrentDOpenID, "--start", "2026-03-10T00:00:00+08:00", "--end", "2026-03-11T00:00:00+08:00", "--limit", "7", "--cursor", "0"},
 	},
 	"chat message send": {
-		"group":     {"chat", "message", "send", "--group", "fixture-conversation", "--text", "hello fixture", "--idempotency-key", "param-alias-equivalence-group", "--yes"},
-		"file-path": {"chat", "message", "send", "--group", "fixture-conversation", "--msg-type", "file", "--file-path", "../../go.mod", "--dentry-id", "1", "--space-id", "2", "--idempotency-key", "param-alias-equivalence-file", "--yes"},
+		"group": {"chat", "message", "send", "--group", "fixture-conversation", "--content", "hello fixture", "--idempotency-key", "param-alias-equivalence-group", "--yes"},
+		"file":  {"chat", "message", "send", "--group", "fixture-conversation", "--msg-type", "file", "--file", "../../go.mod", "--dentry-id", "1", "--space-id", "2", "--idempotency-key", "param-alias-equivalence-file", "--yes"},
 	},
 	"chat +conversation-set-top": {
 		"conversation-ids": {"chat", "+conversation-set-top", "--conversation-ids", "fixture-conversation-1,fixture-conversation-2", "--yes"},
@@ -260,7 +305,6 @@ var paramAliasNewIMCases = []struct {
 	{command: "chat message list-favorites", emitted: "limit", canonical: "size"},
 	{command: "chat message list-unread-conversations", emitted: "limit", canonical: "count"},
 	{command: "chat message list-unread-conversations", emitted: "size", canonical: "count"},
-	{command: "chat message send", emitted: "file", canonical: "file-path"},
 	{command: "chat message send-by-bot", emitted: "at-users", canonical: "at-user-ids"},
 	{command: "chat message send-by-webhook", emitted: "at-user-ids", canonical: "at-users"},
 	{command: "chat +chat-update", emitted: "chat-id", canonical: "group"},
@@ -278,7 +322,7 @@ var paramAliasNewIMCases = []struct {
 	{command: "chat +chat-members-get", emitted: "chat", canonical: "id"},
 	{command: "chat +messages-list", emitted: "start", canonical: "time"},
 	{command: "chat +messages-reply", emitted: "msg-id", canonical: "ref-msg-id"},
-	{command: "chat +messages-reply", emitted: "chat", canonical: "conversation-id"},
+	{command: "chat +messages-reply", emitted: "chat", canonical: "group"},
 	{command: "chat +flag-cancel", emitted: "group", canonical: "conversation-id"},
 	{command: "chat +flag-cancel", emitted: "chat", canonical: "conversation-id"},
 	{command: "chat +flag-create", emitted: "group", canonical: "conversation-id"},
@@ -372,16 +416,113 @@ var paramAliasNewDriveCases = []struct {
 	{command: "drive +upload", emitted: "file-id", canonical: "node"},
 }
 
-// paramAliasNewDriveConfirmationCases selects one newly reviewed alias for
-// every Drive command in the expansion whose declared runtime safety requires
-// confirmation. The full matrix below proves all spellings preserve the
-// confirmed payload; this smaller matrix proves aliases cannot cross the
-// confirmation boundary before any transport call is made.
-var paramAliasNewDriveConfirmationCases = []struct {
+// paramAliasAITableDeleteDisableCompleteCommands contains complete invocations
+// for every AITable delete/disable command whose confirmation boundary is
+// reached by aliases introduced in the AITable expansion. These templates are
+// intentionally separate from paramAliasCompleteCommands: that map mirrors
+// the reviewed validation fixture one-for-one, while this matrix exhaustively
+// proves the safety boundary for generated aliases beyond the fixture sample.
+var paramAliasAITableDeleteDisableCompleteCommands = map[string][]string{
+	"aitable +advperm-disable":      {"aitable", "+advperm-disable", "--base-id", "base-1", "--yes"},
+	"aitable +base-delete":          {"aitable", "+base-delete", "--base-id", "base-1", "--yes"},
+	"aitable +chart-delete":         {"aitable", "+chart-delete", "--base-id", "base-1", "--dashboard-id", "dashboard-1", "--chart-id", "chart-1", "--yes"},
+	"aitable +dashboard-delete":     {"aitable", "+dashboard-delete", "--base-id", "base-1", "--dashboard-id", "dashboard-1", "--yes"},
+	"aitable +field-delete":         {"aitable", "+field-delete", "--base-id", "base-1", "--table-id", "table-1", "--field-id", "field-1", "--yes"},
+	"aitable +form-delete":          {"aitable", "+form-delete", "--base-id", "base-1", "--table-id", "table-1", "--view-id", "view-1", "--yes"},
+	"aitable +record-delete":        {"aitable", "+record-delete", "--base-id", "base-1", "--table-id", "table-1", "--record-ids", "record-1", "--yes"},
+	"aitable +role-delete":          {"aitable", "+role-delete", "--base-id", "base-1", "--role-id", "role-1", "--yes"},
+	"aitable +section-delete":       {"aitable", "+section-delete", "--base-id", "base-1", "--section-id", "section-1", "--yes"},
+	"aitable +table-delete":         {"aitable", "+table-delete", "--base-id", "base-1", "--table-id", "table-1", "--yes"},
+	"aitable +view-delete":          {"aitable", "+view-delete", "--base-id", "base-1", "--table-id", "table-1", "--view-id", "view-1", "--yes"},
+	"aitable +workflow-disable":     {"aitable", "+workflow-disable", "--base-id", "base-1", "--workflow-id", "workflow-1", "--yes"},
+	"aitable advperm disable":       {"aitable", "advperm", "disable", "--base-id", "base-1", "--yes"},
+	"aitable advperm role-delete":   {"aitable", "advperm", "role-delete", "--base-id", "base-1", "--role-id", "role-1", "--yes"},
+	"aitable base delete":           {"aitable", "base", "delete", "--base-id", "base-1", "--yes"},
+	"aitable chart delete":          {"aitable", "chart", "delete", "--base-id", "base-1", "--dashboard-id", "dashboard-1", "--chart-id", "chart-1", "--yes"},
+	"aitable dashboard delete":      {"aitable", "dashboard", "delete", "--base-id", "base-1", "--dashboard-id", "dashboard-1", "--yes"},
+	"aitable field delete":          {"aitable", "field", "delete", "--base-id", "base-1", "--table-id", "table-1", "--field-id", "field-1", "--yes"},
+	"aitable form delete":           {"aitable", "form", "delete", "--base-id", "base-1", "--table-id", "table-1", "--view-id", "view-1", "--yes"},
+	"aitable form questions delete": {"aitable", "form", "questions", "delete", "--base-id", "base-1", "--table-id", "table-1", "--field-id", "field-1", "--yes"},
+	"aitable record delete":         {"aitable", "record", "delete", "--base-id", "base-1", "--table-id", "table-1", "--record-ids", "record-1", "--yes"},
+	"aitable table delete":          {"aitable", "table", "delete", "--base-id", "base-1", "--table-id", "table-1", "--yes"},
+	"aitable view delete":           {"aitable", "view", "delete", "--base-id", "base-1", "--table-id", "table-1", "--view-id", "view-1", "--yes"},
+	"aitable workflow disable":      {"aitable", "workflow", "disable", "--base-id", "base-1", "--workflow-id", "workflow-1", "--yes"},
+}
+
+// paramAliasNewAITableDeleteDisableCases is the exhaustive set of alias
+// tuples newly introduced by this change on AITable delete/disable commands
+// that require confirmation. Every tuple must remain on both sides of the
+// confirmation gate: rejected with zero calls before --yes, and exactly
+// payload-equivalent to its canonical spelling after --yes.
+var paramAliasNewAITableDeleteDisableCases = []struct {
 	command   string
 	emitted   string
 	canonical string
 }{
+	{command: "aitable +advperm-disable", emitted: "base", canonical: "base-id"},
+	{command: "aitable +advperm-disable", emitted: "base-token", canonical: "base-id"},
+	{command: "aitable +base-delete", emitted: "base", canonical: "base-id"},
+	{command: "aitable +base-delete", emitted: "base-token", canonical: "base-id"},
+	{command: "aitable +chart-delete", emitted: "base", canonical: "base-id"},
+	{command: "aitable +chart-delete", emitted: "base-token", canonical: "base-id"},
+	{command: "aitable +dashboard-delete", emitted: "base", canonical: "base-id"},
+	{command: "aitable +dashboard-delete", emitted: "base-token", canonical: "base-id"},
+	{command: "aitable +field-delete", emitted: "base", canonical: "base-id"},
+	{command: "aitable +field-delete", emitted: "base-token", canonical: "base-id"},
+	{command: "aitable +field-delete", emitted: "table", canonical: "table-id"},
+	{command: "aitable +form-delete", emitted: "base", canonical: "base-id"},
+	{command: "aitable +form-delete", emitted: "base-token", canonical: "base-id"},
+	{command: "aitable +form-delete", emitted: "table", canonical: "table-id"},
+	{command: "aitable +record-delete", emitted: "base", canonical: "base-id"},
+	{command: "aitable +record-delete", emitted: "base-token", canonical: "base-id"},
+	{command: "aitable +record-delete", emitted: "table", canonical: "table-id"},
+	{command: "aitable +role-delete", emitted: "base", canonical: "base-id"},
+	{command: "aitable +role-delete", emitted: "base-token", canonical: "base-id"},
+	{command: "aitable +section-delete", emitted: "base", canonical: "base-id"},
+	{command: "aitable +section-delete", emitted: "base-token", canonical: "base-id"},
+	{command: "aitable +table-delete", emitted: "base", canonical: "base-id"},
+	{command: "aitable +table-delete", emitted: "base-token", canonical: "base-id"},
+	{command: "aitable +table-delete", emitted: "table", canonical: "table-id"},
+	{command: "aitable +view-delete", emitted: "base", canonical: "base-id"},
+	{command: "aitable +view-delete", emitted: "base-token", canonical: "base-id"},
+	{command: "aitable +view-delete", emitted: "table", canonical: "table-id"},
+	{command: "aitable +workflow-disable", emitted: "base", canonical: "base-id"},
+	{command: "aitable +workflow-disable", emitted: "base-token", canonical: "base-id"},
+	{command: "aitable +workflow-disable", emitted: "flow-id", canonical: "workflow-id"},
+	{command: "aitable advperm disable", emitted: "base-token", canonical: "base-id"},
+	{command: "aitable advperm role-delete", emitted: "base-token", canonical: "base-id"},
+	{command: "aitable base delete", emitted: "base-token", canonical: "base-id"},
+	{command: "aitable chart delete", emitted: "base-token", canonical: "base-id"},
+	{command: "aitable dashboard delete", emitted: "base-token", canonical: "base-id"},
+	{command: "aitable field delete", emitted: "base-token", canonical: "base-id"},
+	{command: "aitable field delete", emitted: "table", canonical: "table-id"},
+	{command: "aitable form delete", emitted: "base-token", canonical: "base-id"},
+	{command: "aitable form delete", emitted: "table", canonical: "table-id"},
+	{command: "aitable form questions delete", emitted: "base-token", canonical: "base-id"},
+	{command: "aitable form questions delete", emitted: "table", canonical: "table-id"},
+	{command: "aitable record delete", emitted: "base-token", canonical: "base-id"},
+	{command: "aitable record delete", emitted: "table", canonical: "table-id"},
+	{command: "aitable table delete", emitted: "base-token", canonical: "base-id"},
+	{command: "aitable table delete", emitted: "table", canonical: "table-id"},
+	{command: "aitable view delete", emitted: "base-token", canonical: "base-id"},
+	{command: "aitable view delete", emitted: "table", canonical: "table-id"},
+	{command: "aitable workflow disable", emitted: "base-token", canonical: "base-id"},
+	{command: "aitable workflow disable", emitted: "flow-id", canonical: "workflow-id"},
+}
+
+// paramAliasNewConfirmationCases selects newly reviewed aliases for commands
+// whose declared runtime safety requires confirmation. The full matrix below
+// proves all spellings preserve the confirmed payload; this smaller matrix
+// proves aliases cannot cross the confirmation boundary before any transport
+// call is made.
+var paramAliasNewConfirmationCases = []struct {
+	command   string
+	emitted   string
+	canonical string
+}{
+	{command: "aitable workflow run", emitted: "base-token", canonical: "base-id"},
+	{command: "aitable workflow run", emitted: "flow-id", canonical: "workflow-id"},
+	{command: "aitable workflow run", emitted: "table", canonical: "table-id"},
 	{command: "drive +delete", emitted: "file-id", canonical: "node"},
 	{command: "drive +publish-unset", emitted: "document-url", canonical: "node"},
 	{command: "drive +recycle-restore", emitted: "recycle-item-id", canonical: "id"},
@@ -402,7 +543,27 @@ var paramAliasNewDriveConfirmationCases = []struct {
 // That duplicated command construction was enough to push the pre-existing
 // macOS app suite beyond its package-level 10-minute timeout.
 var paramAliasRepresentativePayloadCases = map[string]bool{
+	paramAliasPayloadCaseKey("aitable +export-data", "export-format"):                true, // shortcut-local export format keeps the final payload
+	paramAliasPayloadCaseKey("aitable +find-record", "base-id"):                      true, // shortcut Base ID compatibility
+	paramAliasPayloadCaseKey("aitable +find-record", "table-id"):                     true, // shortcut Table ID compatibility
 	paramAliasPayloadCaseKey("aitable +record-query", "base"):                        true, // concept alias on a shortcut read
+	paramAliasPayloadCaseKey("aitable +record-share-links", "base-id"):               true, // observed experiment Base ID spelling
+	paramAliasPayloadCaseKey("aitable +record-share-links", "table-id"):              true, // observed experiment Table ID spelling
+	paramAliasPayloadCaseKey("aitable +workflow-list", "max-results"):                true, // shortcut pagination-size alias
+	paramAliasPayloadCaseKey("aitable attachment upload", "file-size"):               true, // byte-size command override
+	paramAliasPayloadCaseKey("aitable base list", "next-cursor"):                     true, // cursor concept alias
+	paramAliasPayloadCaseKey("aitable base update", "description"):                   true, // plain description alias on a write command
+	paramAliasPayloadCaseKey("aitable field search-options", "query"):                true, // search keyword concept alias
+	paramAliasPayloadCaseKey("aitable workflow get", "flow-id"):                      true, // workflow ID concept alias
+	paramAliasPayloadCaseKey("aitable workflow history", "base-token"):               true, // Base ID concept alias
+	paramAliasPayloadCaseKey("aitable workflow history", "end-time"):                 true, // upper time-bound override
+	paramAliasPayloadCaseKey("aitable workflow history", "flow-id"):                  true, // workflow ID concept alias
+	paramAliasPayloadCaseKey("aitable workflow history", "page-index"):               true, // zero-based page override
+	paramAliasPayloadCaseKey("aitable workflow history", "page-size"):                true, // page-size concept alias
+	paramAliasPayloadCaseKey("aitable workflow history", "start-time"):               true, // lower time-bound override
+	paramAliasPayloadCaseKey("aitable workflow run", "base-token"):                   true, // Base ID concept alias on a confirmed write
+	paramAliasPayloadCaseKey("aitable workflow run", "flow-id"):                      true, // workflow ID concept alias on a confirmed write
+	paramAliasPayloadCaseKey("aitable workflow run", "table"):                        true, // Table ID concept alias on a confirmed write
 	paramAliasPayloadCaseKey("attendance check result", "user-ids"):                  true, // list-valued concept alias
 	paramAliasPayloadCaseKey("calendar event list", "date"):                          true, // time concept alias
 	paramAliasPayloadCaseKey("chat message add-favorite", "msg-id"):                  true, // scoped IM identifier alias
@@ -426,7 +587,11 @@ var paramAliasRepresentativePayloadCases = map[string]bool{
 	paramAliasPayloadCaseKey("doc +update", "revision"):                              true, // optimistic edit revision alias
 	paramAliasPayloadCaseKey("doc +access-grant", "doc-id"):                          true, // permission write keeps document identity
 	paramAliasPayloadCaseKey("doc +version-revert", "version-number"):                true, // high-write version role with canonical confirmation
-	paramAliasPayloadCaseKey("doc block insert", "content"):                          true, // block write content alias
+	paramAliasPayloadCaseKey("chat +messages-reply", "conversation-id"):              true, // renamed conversation Primary keeps final reply payload
+	paramAliasPayloadCaseKey("chat message send", "file-path"):                       true, // renamed local-file Primary reaches the same final payload
+	paramAliasPayloadCaseKey("doc +doc-append", "text"):                              true, // shortcut content rename keeps append payload
+	paramAliasPayloadCaseKey("doc block insert", "text"):                             true, // block write content compatibility alias
+	paramAliasPayloadCaseKey("doc block update", "text"):                             true, // update uses the same typed compatibility path
 	paramAliasPayloadCaseKey("doc block insert", "parent-block-id"):                  true, // scoped block-role alias
 	paramAliasPayloadCaseKey("doc comment delete", "comment-id"):                     true, // destructive comment-key alias
 	paramAliasPayloadCaseKey("doc comment reply", "mentioned-open-conversation-ids"): true, // list-valued group mention role
@@ -435,6 +600,108 @@ var paramAliasRepresentativePayloadCases = map[string]bool{
 	paramAliasPayloadCaseKey("report list", "from-date"):                             true, // date-range concept alias
 }
 
+// paramAliasCalendarPayloadCases keeps the full reviewed Calendar expansion
+// separate from the long-lived app-c race process. Each case still executes
+// both canonical and alias argv through the real PreParse/Cobra path and
+// compares the final captured transport calls; the owning top-level test runs
+// these allocations in a short-lived race-instrumented subprocess so all Root
+// registrations are released together when that process exits.
+var paramAliasCalendarPayloadCases = map[string]bool{
+	paramAliasPayloadCaseKey("calendar +agenda", "from"):                    true,
+	paramAliasPayloadCaseKey("calendar +agenda", "to"):                      true,
+	paramAliasPayloadCaseKey("calendar +agenda", "max-results"):             true,
+	paramAliasPayloadCaseKey("calendar +agenda", "next-cursor"):             true,
+	paramAliasPayloadCaseKey("calendar +agenda", "calendar-book-id"):        true,
+	paramAliasPayloadCaseKey("calendar +attendee-list", "event-id"):         true,
+	paramAliasPayloadCaseKey("calendar +attendee-list", "calendar-book-id"): true,
+	paramAliasPayloadCaseKey("calendar +book", "summary"):                   true,
+	paramAliasPayloadCaseKey("calendar +book", "attendee-names"):            true,
+	paramAliasPayloadCaseKey("calendar +book-search", "keyword"):            true,
+	paramAliasPayloadCaseKey("calendar +book-search", "search"):             true,
+	paramAliasPayloadCaseKey("calendar +book-search", "name"):               true,
+	paramAliasPayloadCaseKey("calendar +cancel-event", "event-id"):          true,
+	paramAliasPayloadCaseKey("calendar +cancel-event", "id"):                true,
+	paramAliasPayloadCaseKey("calendar +free", "name"):                      true,
+	paramAliasPayloadCaseKey("calendar +free-slots", "start-hour"):          true,
+	paramAliasPayloadCaseKey("calendar +free-slots", "end-hour"):            true,
+	paramAliasPayloadCaseKey("calendar +free-slots", "day-offset"):          true,
+	paramAliasPayloadCaseKey("calendar +freebusy", "user-ids"):              true,
+	paramAliasPayloadCaseKey("calendar +freebusy", "room-ids"):              true,
+	paramAliasPayloadCaseKey("calendar +freebusy", "room-id"):               true,
+	paramAliasPayloadCaseKey("calendar +my-free", "from"):                   true,
+	paramAliasPayloadCaseKey("calendar +my-free", "to"):                     true,
+	paramAliasPayloadCaseKey("calendar +invite", "id"):                      true,
+	paramAliasPayloadCaseKey("calendar +invite", "participant-names"):       true,
+	paramAliasPayloadCaseKey("calendar +reschedule", "id"):                  true,
+	paramAliasPayloadCaseKey("calendar +reschedule", "from"):                true,
+	paramAliasPayloadCaseKey("calendar +reschedule", "to"):                  true,
+	paramAliasPayloadCaseKey("calendar +room-groups", "page-size"):          true,
+	paramAliasPayloadCaseKey("calendar +room-groups", "page-index"):         true,
+	paramAliasPayloadCaseKey("calendar +room-search", "query"):              true,
+	paramAliasPayloadCaseKey("calendar +suggest-time", "duration-minutes"):  true,
+	paramAliasPayloadCaseKey("calendar +suggest-time", "attendee-names"):    true,
+	paramAliasPayloadCaseKey("calendar +conflicts", "day-offset"):           true,
+	paramAliasPayloadCaseKey("calendar busy search", "room-id"):             true,
+	paramAliasPayloadCaseKey("calendar event create", "reminder-minutes"):   true,
+	paramAliasPayloadCaseKey("calendar event create", "tz"):                 true,
+	paramAliasPayloadCaseKey("calendar event create", "room-id"):            true,
+	paramAliasPayloadCaseKey("calendar event respond", "response-status"):   true,
+	paramAliasPayloadCaseKey("calendar event suggest", "duration-minutes"):  true,
+	paramAliasPayloadCaseKey("calendar event update", "tz"):                 true,
+	paramAliasPayloadCaseKey("calendar room add", "room-id"):                true,
+	paramAliasPayloadCaseKey("calendar room delete", "room-id"):             true,
+	paramAliasPayloadCaseKey("calendar room search", "room-group-id"):       true,
+	paramAliasPayloadCaseKey("calendar +create", "summary"):                 true,
+	paramAliasPayloadCaseKey("calendar +create", "description"):             true,
+	paramAliasPayloadCaseKey("calendar +create", "user-ids"):                true,
+	paramAliasPayloadCaseKey("calendar +create", "room-ids"):                true,
+	paramAliasPayloadCaseKey("calendar +create", "room-id"):                 true,
+	paramAliasPayloadCaseKey("calendar +create", "calendar-book-id"):        true,
+	paramAliasPayloadCaseKey("calendar +create", "to"):                      true,
+	paramAliasPayloadCaseKey("calendar +create", "from"):                    true,
+	paramAliasPayloadCaseKey("calendar +get", "event-id"):                   true,
+	paramAliasPayloadCaseKey("calendar +get", "calendar-book-id"):           true,
+	paramAliasPayloadCaseKey("calendar +room-find", "from"):                 true,
+	paramAliasPayloadCaseKey("calendar +room-find", "to"):                   true,
+	paramAliasPayloadCaseKey("calendar +room-find", "page-size"):            true,
+	paramAliasPayloadCaseKey("calendar +room-find", "page-index"):           true,
+	paramAliasPayloadCaseKey("calendar +room-find", "room-group-id"):        true,
+	paramAliasPayloadCaseKey("calendar +room-find", "query"):                true,
+	paramAliasPayloadCaseKey("calendar +rsvp", "event-id"):                  true,
+	paramAliasPayloadCaseKey("calendar +rsvp", "response-status"):           true,
+	paramAliasPayloadCaseKey("calendar +search-event", "keyword"):           true,
+	paramAliasPayloadCaseKey("calendar +search-event", "from"):              true,
+	paramAliasPayloadCaseKey("calendar +search-event", "to"):                true,
+	paramAliasPayloadCaseKey("calendar +search-event", "next-cursor"):       true,
+	paramAliasPayloadCaseKey("calendar +search-event", "max-results"):       true,
+	paramAliasPayloadCaseKey("calendar +suggestion", "user-ids"):            true,
+	paramAliasPayloadCaseKey("calendar +suggestion", "duration-minutes"):    true,
+	paramAliasPayloadCaseKey("calendar +suggestion", "from"):                true,
+	paramAliasPayloadCaseKey("calendar +suggestion", "to"):                  true,
+	paramAliasPayloadCaseKey("calendar +suggestion", "tz"):                  true,
+	paramAliasPayloadCaseKey("calendar +update", "event-id"):                true,
+	paramAliasPayloadCaseKey("calendar +update", "from"):                    true,
+	paramAliasPayloadCaseKey("calendar +update", "summary"):                 true,
+	paramAliasPayloadCaseKey("calendar +update", "description"):             true,
+	paramAliasPayloadCaseKey("calendar +update", "add-user-ids"):            true,
+	paramAliasPayloadCaseKey("calendar +update", "remove-user-ids"):         true,
+}
+
+// paramAliasCalendarConfirmationCases selects one newly reviewed alias for
+// every Calendar Shortcut whose runtime contract requires user confirmation.
+// The complete Calendar matrix proves confirmed canonical/alias payload
+// equality; these representatives additionally prove semantic normalization
+// cannot cross the confirmation boundary before the first transport call.
+var paramAliasCalendarConfirmationCases = map[string]bool{
+	paramAliasPayloadCaseKey("calendar +book", "summary"):          true,
+	paramAliasPayloadCaseKey("calendar +cancel-event", "event-id"): true,
+	paramAliasPayloadCaseKey("calendar +create", "summary"):        true,
+	paramAliasPayloadCaseKey("calendar +invite", "id"):             true,
+	paramAliasPayloadCaseKey("calendar +reschedule", "from"):       true,
+	paramAliasPayloadCaseKey("calendar +rsvp", "response-status"):  true,
+	paramAliasPayloadCaseKey("calendar +update", "event-id"):       true,
+}
+
 func TestCrossPlatformCoverageReviewedParamAliasesHaveCompleteTemplatesAndRepresentativeFinalPayloads(t *testing.T) {
 	concepts, err := cli.LoadParamConcepts()
 	if err != nil {
@@ -468,28 +735,7 @@ func TestCrossPlatformCoverageReviewedParamAliasesHaveCompleteTemplatesAndRepres
 		}
 		executedRepresentatives[caseKey] = true
 		t.Run(fixture.Command+"/"+fixture.Emitted, func(t *testing.T) {
-
-			canonicalCaller := &paramAliasCaptureCaller{}
-			_, canonicalErr := executeParamAliasPayloadE2E(t, canonicalCaller, canonicalArgs...)
-			if canonicalErr != nil {
-				t.Fatalf("complete canonical command failed: %v\nargs=%v\ncalls=%#v", canonicalErr, canonicalArgs, canonicalCaller.calls)
-			}
-			if len(canonicalCaller.calls) == 0 {
-				t.Fatalf("complete canonical command reached no final transport payload: args=%v", canonicalArgs)
-			}
-
-			aliasCaller := &paramAliasCaptureCaller{}
-			ctx, aliasErr := executeParamAliasPayloadE2E(t, aliasCaller, aliasArgs...)
-			if aliasErr != nil {
-				t.Fatalf("complete alias command failed: %v\nargs=%v\ncalls=%#v", aliasErr, aliasArgs, aliasCaller.calls)
-			}
-			if ctx == nil {
-				t.Fatal("complete alias command skipped PreParse")
-			}
-			normalizeParamAliasVolatileDefaults(fixture.Command, canonicalCaller, aliasCaller)
-			if !reflect.DeepEqual(aliasCaller.calls, canonicalCaller.calls) {
-				t.Fatalf("final transport calls differ\ncanonical args: %v\nalias args: %v\ncanonical calls: %#v\nalias calls: %#v", canonicalArgs, aliasArgs, canonicalCaller.calls, aliasCaller.calls)
-			}
+			assertParamAliasFinalPayloadEquivalent(t, fixture.Command, canonicalArgs, aliasArgs)
 		})
 	}
 
@@ -519,6 +765,118 @@ func TestCrossPlatformCoverageReviewedParamAliasesHaveCompleteTemplatesAndRepres
 	}
 }
 
+func TestCrossPlatformCoverageReviewedCalendarParamAliasesReachCanonicalEquivalentFinalPayloads(t *testing.T) {
+	if os.Getenv(paramAliasCalendarPayloadChildEnv) != "1" {
+		command := exec.Command(
+			os.Args[0],
+			"-test.run=^TestCrossPlatformCoverageReviewedCalendarParamAliasesReachCanonicalEquivalentFinalPayloads$",
+			"-test.count=1",
+			"-test.timeout=5m",
+		)
+		command.Env = append(os.Environ(), paramAliasCalendarPayloadChildEnv+"=1")
+		output, err := command.CombinedOutput()
+		if err != nil {
+			t.Fatalf("Calendar param-alias payload subprocess failed: %v\n%s", err, strings.TrimSpace(string(output)))
+		}
+		return
+	}
+
+	concepts, err := cli.LoadParamConcepts()
+	if err != nil {
+		t.Fatalf("LoadParamConcepts() error = %v", err)
+	}
+
+	executed := make(map[string]bool)
+	executedConfirmation := make(map[string]bool)
+	for _, fixture := range concepts.Fixture {
+		caseKey := paramAliasPayloadCaseKey(fixture.Command, fixture.Emitted)
+		if !paramAliasCalendarPayloadCases[caseKey] {
+			continue
+		}
+		executed[caseKey] = true
+		fixture := fixture
+		t.Run(fixture.Command+"/"+fixture.Emitted, func(t *testing.T) {
+			complete, ok := paramAliasCompleteCommand(fixture.Command, fixture.Expect)
+			if !ok {
+				t.Fatal("reviewed Calendar alias has no complete-command E2E template")
+			}
+			canonicalArgs := append([]string(nil), complete...)
+			aliasArgs, replacements := replaceLongFlag(canonicalArgs, fixture.Expect, fixture.Emitted)
+			if replacements != 1 {
+				t.Fatalf("complete Calendar command must contain canonical --%s exactly once; replacements=%d args=%v", fixture.Expect, replacements, canonicalArgs)
+			}
+			assertParamAliasFinalPayloadEquivalent(t, fixture.Command, canonicalArgs, aliasArgs)
+			if paramAliasCalendarConfirmationCases[caseKey] {
+				executedConfirmation[caseKey] = true
+				assertParamAliasCannotBypassConfirmation(t, aliasArgs)
+			}
+		})
+	}
+
+	for caseKey := range paramAliasCalendarPayloadCases {
+		if !executed[caseKey] {
+			t.Errorf("Calendar final-payload case %q has no active reviewed fixture", caseKey)
+		}
+	}
+	if len(executed) != len(paramAliasCalendarPayloadCases) {
+		t.Fatalf("Calendar final-payload coverage = %d, want %d", len(executed), len(paramAliasCalendarPayloadCases))
+	}
+	for caseKey := range paramAliasCalendarConfirmationCases {
+		if !executedConfirmation[caseKey] {
+			t.Errorf("Calendar confirmation case %q has no active reviewed fixture", caseKey)
+		}
+	}
+	if len(executedConfirmation) != len(paramAliasCalendarConfirmationCases) {
+		t.Fatalf("Calendar confirmation coverage = %d, want %d", len(executedConfirmation), len(paramAliasCalendarConfirmationCases))
+	}
+}
+
+func assertParamAliasCannotBypassConfirmation(t *testing.T, aliasArgs []string) {
+	t.Helper()
+	unconfirmedArgs, removals := removeExactArg(aliasArgs, "--yes")
+	if removals != 1 {
+		t.Fatalf("confirmation template must contain --yes exactly once; removals=%d args=%v", removals, aliasArgs)
+	}
+
+	caller := &paramAliasCaptureCaller{}
+	ctx, err := executeParamAliasPayloadE2E(t, caller, unconfirmedArgs...)
+	if ctx == nil {
+		t.Fatal("unconfirmed Calendar alias command skipped PreParse")
+	}
+	var appErr *apperrors.Error
+	if !errors.As(err, &appErr) || appErr.Reason != "confirmation_required" {
+		t.Fatalf("unconfirmed Calendar alias command error = %#v, want confirmation_required\nargs=%v", err, unconfirmedArgs)
+	}
+	if len(caller.calls) != 0 {
+		t.Fatalf("unconfirmed Calendar alias crossed the transport boundary: args=%v calls=%#v", unconfirmedArgs, caller.calls)
+	}
+}
+
+func assertParamAliasFinalPayloadEquivalent(t *testing.T, command string, canonicalArgs, aliasArgs []string) {
+	t.Helper()
+	canonicalCaller := &paramAliasCaptureCaller{}
+	_, canonicalErr := executeParamAliasPayloadE2E(t, canonicalCaller, canonicalArgs...)
+	if canonicalErr != nil {
+		t.Fatalf("complete canonical command failed: %v\nargs=%v\ncalls=%#v", canonicalErr, canonicalArgs, canonicalCaller.calls)
+	}
+	if len(canonicalCaller.calls) == 0 {
+		t.Fatalf("complete canonical command reached no final transport payload: args=%v", canonicalArgs)
+	}
+
+	aliasCaller := &paramAliasCaptureCaller{}
+	ctx, aliasErr := executeParamAliasPayloadE2E(t, aliasCaller, aliasArgs...)
+	if aliasErr != nil {
+		t.Fatalf("complete alias command failed: %v\nargs=%v\ncalls=%#v", aliasErr, aliasArgs, aliasCaller.calls)
+	}
+	if ctx == nil {
+		t.Fatal("complete alias command skipped PreParse")
+	}
+	normalizeParamAliasVolatileDefaults(command, canonicalCaller, aliasCaller)
+	if !reflect.DeepEqual(aliasCaller.calls, canonicalCaller.calls) {
+		t.Fatalf("final transport calls differ\ncanonical args: %v\nalias args: %v\ncanonical calls: %#v\nalias calls: %#v", canonicalArgs, aliasArgs, canonicalCaller.calls, aliasCaller.calls)
+	}
+}
+
 func TestCrossPlatformCoverageNewIMParamAliasesReachCanonicalEquivalentFinalPayloads(t *testing.T) {
 	activeAliases := 0
 	for _, test := range paramAliasNewIMCases {
@@ -629,13 +987,116 @@ func TestCrossPlatformCoverageNewDriveParamAliasesReachCanonicalEquivalentFinalP
 	}
 }
 
-func TestCrossPlatformCoverageNewDriveParamAliasesCannotBypassConfirmation(t *testing.T) {
-	for _, test := range paramAliasNewDriveConfirmationCases {
+func TestCrossPlatformCoverageNewAITableDeleteDisableAliasesPreserveConfirmationAndPayload(t *testing.T) {
+	coveredCommands := make(map[string]bool)
+	reviewedAliases := make(map[string]string, len(paramAliasNewAITableDeleteDisableCases))
+	for _, test := range paramAliasNewAITableDeleteDisableCases {
+		test := test
+		caseKey := paramAliasPayloadCaseKey(test.command, test.emitted)
+		if previous, duplicate := reviewedAliases[caseKey]; duplicate {
+			t.Fatalf("duplicate AITable delete/disable alias case %q: --%s and --%s", caseKey, previous, test.canonical)
+		}
+		reviewedAliases[caseKey] = test.canonical
+		t.Run(test.command+"/"+test.emitted, func(t *testing.T) {
+			complete, ok := paramAliasAITableDeleteDisableCompleteCommands[test.command]
+			if !ok {
+				t.Fatal("reviewed AITable delete/disable alias has no complete safety template")
+			}
+			coveredCommands[test.command] = true
+
+			canonicalArgs := append([]string(nil), complete...)
+			aliasArgs, replacements := replaceLongFlag(canonicalArgs, test.canonical, test.emitted)
+			if replacements != 1 {
+				t.Fatalf("complete command must contain canonical --%s exactly once; replacements=%d args=%v", test.canonical, replacements, canonicalArgs)
+			}
+			unconfirmedArgs, removals := removeExactArg(aliasArgs, "--yes")
+			if removals != 1 {
+				t.Fatalf("safety template must contain --yes exactly once; removals=%d args=%v", removals, aliasArgs)
+			}
+
+			entry, exists := cli.LookupParamAlias(test.command)
+			target, active := entry.ResolveAlias(test.emitted)
+			if !exists || !active || target != test.canonical {
+				t.Fatalf("reviewed AITable alias --%s resolution = exists:%v active:%v target:%q, want --%s", test.emitted, exists, active, target, test.canonical)
+			}
+
+			unconfirmedCaller := &paramAliasCaptureCaller{}
+			ctx, unconfirmedErr := executeParamAliasPayloadE2E(t, unconfirmedCaller, unconfirmedArgs...)
+			if ctx == nil {
+				t.Fatal("unconfirmed AITable alias command skipped PreParse")
+			}
+			var appErr *apperrors.Error
+			if !errors.As(unconfirmedErr, &appErr) || appErr.Reason != "confirmation_required" {
+				t.Fatalf("unconfirmed AITable alias command error = %#v, want confirmation_required\nargs=%v", unconfirmedErr, unconfirmedArgs)
+			}
+			if len(unconfirmedCaller.calls) != 0 {
+				t.Fatalf("unconfirmed AITable alias crossed the transport boundary: args=%v calls=%#v", unconfirmedArgs, unconfirmedCaller.calls)
+			}
+
+			canonicalCaller := &paramAliasCaptureCaller{}
+			_, canonicalErr := executeParamAliasPayloadE2E(t, canonicalCaller, canonicalArgs...)
+			if canonicalErr != nil {
+				t.Fatalf("confirmed canonical command failed: %v\nargs=%v\ncalls=%#v", canonicalErr, canonicalArgs, canonicalCaller.calls)
+			}
+			if len(canonicalCaller.calls) == 0 {
+				t.Fatalf("confirmed canonical command reached no final transport payload: args=%v", canonicalArgs)
+			}
+
+			aliasCaller := &paramAliasCaptureCaller{}
+			aliasCtx, aliasErr := executeParamAliasPayloadE2E(t, aliasCaller, aliasArgs...)
+			if aliasErr != nil {
+				t.Fatalf("confirmed alias command failed: %v\nargs=%v\ncalls=%#v", aliasErr, aliasArgs, aliasCaller.calls)
+			}
+			if aliasCtx == nil {
+				t.Fatal("confirmed AITable alias command skipped PreParse")
+			}
+			if !reflect.DeepEqual(aliasCaller.calls, canonicalCaller.calls) {
+				t.Fatalf("confirmed final transport calls differ\ncanonical args: %v\nalias args: %v\ncanonical calls: %#v\nalias calls: %#v", canonicalArgs, aliasArgs, canonicalCaller.calls, aliasCaller.calls)
+			}
+		})
+	}
+
+	for command := range paramAliasAITableDeleteDisableCompleteCommands {
+		if !coveredCommands[command] {
+			t.Errorf("AITable delete/disable safety template %q has no reviewed alias case", command)
+		}
+	}
+	if len(coveredCommands) != len(paramAliasAITableDeleteDisableCompleteCommands) {
+		t.Fatalf("AITable delete/disable safety coverage = %d commands, want %d", len(coveredCommands), len(paramAliasAITableDeleteDisableCompleteCommands))
+	}
+
+	activeAliases := 0
+	for command := range paramAliasAITableDeleteDisableCompleteCommands {
+		entry, exists := cli.LookupParamAlias(command)
+		if !exists {
+			t.Errorf("AITable delete/disable safety command %q has no generated alias entry", command)
+			continue
+		}
+		for emitted, canonical := range entry.Aliases {
+			activeAliases++
+			caseKey := paramAliasPayloadCaseKey(command, emitted)
+			reviewedCanonical, reviewed := reviewedAliases[caseKey]
+			if !reviewed {
+				t.Errorf("active AITable delete/disable alias %q --%s -> --%s has no confirmation/payload case", command, emitted, canonical)
+				continue
+			}
+			if reviewedCanonical != canonical {
+				t.Errorf("reviewed AITable delete/disable alias %q --%s target = --%s, generated --%s", command, emitted, reviewedCanonical, canonical)
+			}
+		}
+	}
+	if activeAliases != len(reviewedAliases) {
+		t.Fatalf("AITable delete/disable generated alias coverage = %d, want %d reviewed cases", activeAliases, len(reviewedAliases))
+	}
+}
+
+func TestCrossPlatformCoverageNewParamAliasesCannotBypassConfirmation(t *testing.T) {
+	for _, test := range paramAliasNewConfirmationCases {
 		test := test
 		t.Run(test.command+"/"+test.emitted, func(t *testing.T) {
 			complete, ok := paramAliasCompleteCommand(test.command, test.canonical)
 			if !ok {
-				t.Fatal("reviewed Drive confirmation alias has no complete-command E2E template")
+				t.Fatal("reviewed confirmation alias has no complete-command E2E template")
 			}
 			aliasArgs, replacements := replaceLongFlag(complete, test.canonical, test.emitted)
 			if replacements != 1 {
@@ -649,7 +1110,7 @@ func TestCrossPlatformCoverageNewDriveParamAliasesCannotBypassConfirmation(t *te
 			entry, exists := cli.LookupParamAlias(test.command)
 			target, active := entry.ResolveAlias(test.emitted)
 			if !exists || !active || target != test.canonical {
-				t.Fatalf("reviewed Drive alias --%s resolution = exists:%v active:%v target:%q, want --%s", test.emitted, exists, active, target, test.canonical)
+				t.Fatalf("reviewed confirmation alias --%s resolution = exists:%v active:%v target:%q, want --%s", test.emitted, exists, active, target, test.canonical)
 			}
 
 			caller := &paramAliasCaptureCaller{}

internal/app/profile_args_test.go

@@ -66,6 +66,32 @@ func TestPreparseProfileFlagUsesNormalizedProfileArgs(t *testing.T) {
 	}
 }
 
+func TestCrossPlatformCoveragePreparseProfileFlagUsesLastOccurrence(t *testing.T) {
+	for _, tc := range []struct {
+		name  string
+		args  []string
+		want  string
+		valid bool
+	}{
+		{name: "space then equals", args: []string{"--profile", "corp-a", "version", "--profile=corp-b"}, want: "corp-b", valid: true},
+		{name: "equals then space", args: []string{"--profile=corp-a", "version", "--profile", "corp-b"}, want: "corp-b", valid: true},
+		{name: "last multi", args: []string{"--profile=corp-a", "--profile", "corp-b,", "corp-c", "version"}, want: "corp-b,corp-c", valid: true},
+		{name: "empty equals clears earlier", args: []string{"--profile=corp-a", "version", "--profile="}},
+		{name: "missing value clears earlier", args: []string{"--profile=corp-a", "version", "--profile"}},
+		{name: "next flag is not profile value", args: []string{"--profile=corp-a", "--profile", "--debug", "version"}},
+	} {
+		t.Run(tc.name, func(t *testing.T) {
+			if got := preparseProfileFlag(tc.args); got != tc.want {
+				t.Fatalf("preparseProfileFlag(%#v) = %q, want %q", tc.args, got, tc.want)
+			}
+			_, specified, valid := preparseProfileSelection(tc.args)
+			if !specified || valid != tc.valid {
+				t.Fatalf("preparseProfileSelection(%#v) = specified %v valid %v, want true/%v", tc.args, specified, valid, tc.valid)
+			}
+		})
+	}
+}
+
 func TestNormalizeProcessProfileArgsRestoresOriginalArgv(t *testing.T) {
 	oldArgs := os.Args
 	t.Cleanup(func() { os.Args = oldArgs })

internal/app/root.go

@@ -80,10 +80,19 @@ var (
 	rootAuthLoadTokenData           = authpkg.LoadTokenData
 	rootNewCommandRunnerWithFlags   = newCommandRunnerWithFlags
 	rootEmitResult                  = output.EmitResult
+	rootInstallProcessSignalContext = installProcessSignalContext
 )
 
 // Execute runs the root command and returns the process exit code.
-func Execute() (exitCode int) {
+func Execute() int {
+	exitCode, _, _ := ExecuteWithTelemetry()
+	return exitCode
+}
+
+// ExecuteWithTelemetry runs the root command and additionally returns a
+// privacy-safe command path and error summary for the official CLI entrypoint.
+func ExecuteWithTelemetry() (exitCode int, commandPath string, errorMessage string) {
+	commandPath = "dws"
 	var (
 		root        *cobra.Command
 		executed    *cobra.Command
@@ -91,12 +100,16 @@ func Execute() (exitCode int) {
 	)
 	defer func() {
 		if r := recover(); r != nil {
+			errorMessage = "internal panic"
 			target := executed
 			if target == nil && root != nil {
 				if found, _, err := root.Find(os.Args[1:]); err == nil {
 					target = found
 				}
 			}
+			if target != nil {
+				commandPath = telemetryCommandPath(target)
+			}
 			if code, attempted, _, _ := output.StoredEmissionState(resultStore); attempted {
 				exitCode = code
 				if target != nil {
@@ -121,6 +134,7 @@ func Execute() (exitCode int) {
 		CloseFileLogger()
 		if executed != nil {
 			if err := closeOutputSink(executed); err != nil {
+				errorMessage = telemetryErrorSummary(err)
 				if code, handled, emitErr := emitOutputPublicationFailure(executed, err); handled && emitErr == nil {
 					exitCode = code
 				} else {
@@ -144,7 +158,9 @@ func Execute() (exitCode int) {
 	agentMetadata := readAgentMetadataSnapshot()
 	if err := agentMetadata.validationError(); err != nil {
 		emitEarlyAgentMetadataValidationError(err, os.Args[1:])
-		return apperrors.ExitCode(err)
+		errorMessage = telemetryErrorSummary(err)
+		exitCode = apperrors.ExitCode(err)
+		return
 	}
 
 	timing := NewTimingCollector()
@@ -162,12 +178,13 @@ func Execute() (exitCode int) {
 	ctx, resultStore = output.WithResultStore(ctx)
 	var signalState *processSignalState
 	var stopSignals func()
-	ctx, signalState, stopSignals = installProcessSignalContext(ctx, resultStore)
+	ctx, signalState, stopSignals = rootInstallProcessSignalContext(ctx, resultStore)
 	defer stopSignals()
 
 	initStart := time.Now()
 	engine := newPipelineEngine()
 	root = rootNewRootCommandWithEngine(ctx, engine)
+	commandPath = telemetryCommandPath(root)
 	timing.Record("cmd_init", time.Since(initStart))
 
 	// Run PreParse handlers on raw argv before Cobra parses flags.
@@ -182,15 +199,23 @@ func Execute() (exitCode int) {
 			result := output.FailureWithExitCode(errorInfoFromExecutionError(err), apperrors.ExitCode(err))
 			code, emitErr := output.EmitResult(target, result)
 			if emitErr == nil {
-				return code
+				errorMessage = telemetryErrorSummary(err)
+				exitCode = code
+				return
 			}
 		}
 		_ = printExecutionError(root, os.Stdout, os.Stderr, err)
-		return apperrors.ExitCode(err)
+		errorMessage = telemetryErrorSummary(err)
+		exitCode = apperrors.ExitCode(err)
+		return
 	}
+	commandPath = telemetryCommandPathForArgs(root, os.Args[1:])
 
 	var err error
 	executed, err = rootExecuteCommand(root)
+	if executed != nil {
+		commandPath = telemetryCommandPath(executed)
+	}
 	// PersistentPostRunE normally commits or aborts the transactional output
 	// sink. Finalize once more at the process boundary so custom execution
 	// seams, embedding callers, or future hook changes cannot leave publication
@@ -222,7 +247,9 @@ func Execute() (exitCode int) {
 				// successfully emitted result into a contradictory 130/143 process
 				// status; likewise, a failed publication must retain its internal
 				// error code instead of being relabelled as cancellation.
-				return code
+				errorMessage = telemetryErrorSummary(interrupted)
+				exitCode = code
+				return
 			}
 		}
 		var publicationErr *outputPublicationError
@@ -233,20 +260,26 @@ func Execute() (exitCode int) {
 	if err != nil {
 		if executed == nil {
 			executed = root
+			commandPath = telemetryCommandPath(root)
 		}
 		if code, attempted, _, _ := output.StoredEmissionState(resultStore); attempted {
 			var publicationErr *outputPublicationError
 			if stderrors.As(err, &publicationErr) {
+				errorMessage = telemetryErrorSummary(publicationErr)
 				if failureCode, handled, emitErr := emitOutputPublicationFailure(executed, publicationErr); handled {
 					if emitErr == nil {
-						return failureCode
+						exitCode = failureCode
+						return
 					}
 					fmt.Fprintf(executed.ErrOrStderr(), "Warning: emit output publication failure: %v\n", emitErr)
 				}
-				return apperrors.ExitCode(publicationErr)
+				exitCode = apperrors.ExitCode(publicationErr)
+				return
 			}
 			fmt.Fprintf(executed.ErrOrStderr(), "Warning: command hook failed after result emission: %v\n", err)
-			return code
+			errorMessage = telemetryErrorSummary(err)
+			exitCode = code
+			return
 		}
 		err = rewordRequiredFlagError(err)
 		var raw apperrors.RawStderrError
@@ -254,7 +287,9 @@ func Execute() (exitCode int) {
 			result := output.FailureWithExitCode(errorInfoFromExecutionError(err), apperrors.ExitCode(err))
 			code, emitErr := output.EmitResult(executed, result)
 			if emitErr == nil {
-				return code
+				errorMessage = telemetryErrorSummary(err)
+				exitCode = code
+				return
 			}
 			err = apperrors.NewInternal("emit failure result: "+emitErr.Error(), apperrors.WithCause(emitErr))
 		}
@@ -264,12 +299,42 @@ func Execute() (exitCode int) {
 			_, _ = fmt.Fprintln(os.Stderr)
 		}
 		_ = printExecutionError(executed, os.Stdout, os.Stderr, err)
-		return apperrors.ExitCode(err)
+		errorMessage = telemetryErrorSummary(err)
+		exitCode = apperrors.ExitCode(err)
+		return
 	}
 	if code, emitted := output.StoredExitCode(resultStore); emitted {
-		return code
+		exitCode = code
+		return
 	}
-	return 0
+	return
+}
+
+func telemetryCommandPath(command *cobra.Command) string {
+	if command == nil {
+		return "dws"
+	}
+	path := strings.TrimSpace(command.CommandPath())
+	root := command.Root()
+	rootName := strings.TrimSpace(root.Name())
+	if path == rootName {
+		return rootName
+	}
+	if rootName != "" {
+		path = strings.TrimSpace(strings.TrimPrefix(path, rootName+" "))
+	}
+	return path
+}
+
+func telemetryCommandPathForArgs(root *cobra.Command, args []string) string {
+	if root == nil {
+		return "dws"
+	}
+	command, _, err := root.Find(args)
+	if err != nil || command == nil {
+		return telemetryCommandPath(root)
+	}
+	return telemetryCommandPath(command)
 }
 
 // emitEarlyAgentMetadataValidationError preserves each built-in command's
@@ -511,6 +576,22 @@ func flagErrorWithSuggestions(cmd *cobra.Command, err error) error {
 	// 无论哪种格式，子串 "--help' for usage." 都可被检索到。
 	tail := fmt.Sprintf("\nSee '%s --help' for usage.", cmd.CommandPath())
 	msgWithTail := errMsg + tail
+	if flag, ok := unknownFlagName(errMsg); ok && flag == "from" {
+		switch cmd.CommandPath() {
+		case "dws chat +search-msg", "dws chat +chat-messages":
+			return apperrors.NewValidation(
+				msgWithTail,
+				apperrors.WithHint("--from 在消息查询中含义不明确：按发送者过滤请使用 --sender <姓名|userId|openDingTalkId>；指定时间起点请使用 --start <RFC3339>"),
+				apperrors.WithReason("ambiguous_flag"),
+				apperrors.WithCause(err),
+				apperrors.WithActions(
+					"Use --sender <姓名|userId|openDingTalkId> to filter by sender",
+					"Use --start <RFC3339> together with --end <RFC3339> to set a time range",
+				),
+				apperrors.WithAvailableFlags(cmdutil.VisibleFlagNames(cmd)...),
+			)
+		}
+	}
 	if flag, protection, ok := reviewedFlagProtection(cmd, errMsg); ok {
 		hint := fmt.Sprintf("Parameter --%s is blocked from automatic normalization on %q; choose an explicit flag from --help.", flag, cmd.CommandPath())
 		reason := "blocked_flag"
@@ -579,15 +660,10 @@ func reviewedFlagProtection(cmd *cobra.Command, errMsg string) (string, pipeline
 	if cmd == nil {
 		return "", "", false
 	}
-	const prefix = "unknown flag: --"
-	idx := strings.Index(errMsg, prefix)
-	if idx < 0 {
+	flag, ok := unknownFlagName(errMsg)
+	if !ok {
 		return "", "", false
 	}
-	flag := strings.TrimSpace(errMsg[idx+len(prefix):])
-	if i := strings.IndexAny(flag, " =\n\t"); i >= 0 {
-		flag = flag[:i]
-	}
 	entry, ok := cli.LookupParamAlias(cmd.CommandPath())
 	if !ok {
 		return "", "", false
@@ -602,6 +678,19 @@ func reviewedFlagProtection(cmd *cobra.Command, errMsg string) (string, pipeline
 	return "", "", false
 }
 
+func unknownFlagName(errMsg string) (string, bool) {
+	const prefix = "unknown flag: --"
+	idx := strings.Index(errMsg, prefix)
+	if idx < 0 {
+		return "", false
+	}
+	flag := strings.TrimSpace(errMsg[idx+len(prefix):])
+	if i := strings.IndexAny(flag, " =\n\t"); i >= 0 {
+		flag = flag[:i]
+	}
+	return flag, flag != ""
+}
+
 func printExecutionError(root *cobra.Command, stdout, stderr io.Writer, err error) error {
 	var raw apperrors.RawStderrError
 	if stderrors.As(err, &raw) {
@@ -935,17 +1024,36 @@ func installReviewedFlagProtectionHandlers(root *cobra.Command) {
 }
 
 func preparseProfileFlag(args []string) string {
+	profile, _, valid := preparseProfileSelection(args)
+	if !valid {
+		return ""
+	}
+	return profile
+}
+
+func preparseProfileSelection(args []string) (profile string, specified, valid bool) {
 	args, _ = normalizeProfileFlagArgs(args)
+	valid = true
 	for i := 0; i < len(args); i++ {
 		arg := strings.TrimSpace(args[i])
 		switch {
-		case arg == "--profile" && i+1 < len(args):
-			return strings.TrimSpace(args[i+1])
+		case arg == "--profile":
+			specified = true
+			if i+1 >= len(args) || strings.HasPrefix(strings.TrimSpace(args[i+1]), "-") {
+				profile = ""
+				valid = false
+				continue
+			}
+			profile = strings.TrimSpace(args[i+1])
+			valid = profile != ""
+			i++
 		case strings.HasPrefix(arg, "--profile="):
-			return strings.TrimSpace(strings.TrimPrefix(arg, "--profile="))
+			specified = true
+			profile = strings.TrimSpace(strings.TrimPrefix(arg, "--profile="))
+			valid = profile != ""
 		}
 	}
-	return ""
+	return profile, specified, valid
 }
 
 func normalizeProcessProfileArgs() func() {

internal/app/root_full_coverage_test.go

@@ -37,29 +37,64 @@ func TestCrossPlatformCoverageRootExecuteAllBranchesCoverage(t *testing.T) {
 	rootNormalizeProcessProfileArgs = func() func() { return func() {} }
 	rootRunPreParse = func(*cobra.Command, *pipeline.Engine) error { return nil }
 	rootStopAllStdioClients = func() {}
+	var executedLeaf *cobra.Command
 	rootNewRootCommandWithEngine = func(context.Context, *pipeline.Engine) *cobra.Command {
-		return &cobra.Command{Use: "dws", SilenceErrors: true, SilenceUsage: true}
+		root := &cobra.Command{Use: "dws", SilenceErrors: true, SilenceUsage: true}
+		sheet := &cobra.Command{Use: "sheet"}
+		executedLeaf = &cobra.Command{Use: "read", Run: func(*cobra.Command, []string) {}}
+		sheet.AddCommand(executedLeaf)
+		root.AddCommand(sheet)
+		return root
 	}
-	rootExecuteCommand = func(cmd *cobra.Command) (*cobra.Command, error) { return cmd, nil }
-	if code := Execute(); code != 0 {
-		t.Fatalf("successful Execute code = %d", code)
+	rootExecuteCommand = func(*cobra.Command) (*cobra.Command, error) { return executedLeaf, nil }
+	if code, commandPath, errorMessage := ExecuteWithTelemetry(); code != 0 || commandPath != "sheet read" || errorMessage != "" {
+		t.Fatalf("successful ExecuteWithTelemetry = code %d path %q error %q", code, commandPath, errorMessage)
 	}
 
 	rootRunPreParse = func(*cobra.Command, *pipeline.Engine) error { return errors.New("alias/canonical conflict") }
-	if code := Execute(); code == 0 {
-		t.Fatal("pre-parse conflict returned zero")
+	if code, _, errorMessage := ExecuteWithTelemetry(); code == 0 || errorMessage != "alias/canonical conflict" {
+		t.Fatalf("pre-parse conflict = code %d error %q", code, errorMessage)
 	}
 	rootRunPreParse = func(*cobra.Command, *pipeline.Engine) error { return nil }
 
 	wantErr := errors.New("unknown command missing")
 	rootExecuteCommand = func(*cobra.Command) (*cobra.Command, error) { return nil, wantErr }
-	if code := Execute(); code == 0 {
-		t.Fatal("failed Execute returned zero")
+	if code, _, errorMessage := ExecuteWithTelemetry(); code == 0 || errorMessage != "unknown command" {
+		t.Fatalf("failed ExecuteWithTelemetry = code %d error %q", code, errorMessage)
 	}
 
 	rootExecuteCommand = func(*cobra.Command) (*cobra.Command, error) { panic("boom") }
-	if code := Execute(); code != 5 {
-		t.Fatalf("panic Execute code = %d", code)
+	os.Args = []string{"dws", "sheet", "read"}
+	if code, commandPath, errorMessage := ExecuteWithTelemetry(); code != 5 || commandPath != "sheet read" || errorMessage != "internal panic" {
+		t.Fatalf("panic ExecuteWithTelemetry = code %d path %q error %q", code, commandPath, errorMessage)
+	}
+}
+
+func TestCrossPlatformCoverageTelemetryCommandPath(t *testing.T) {
+	if got := telemetryCommandPath(nil); got != "dws" {
+		t.Fatalf("nil command path = %q, want dws", got)
+	}
+	if got := telemetryCommandPathForArgs(nil, nil); got != "dws" {
+		t.Fatalf("nil root command path = %q, want dws", got)
+	}
+	root := &cobra.Command{Use: "dws"}
+	sheet := &cobra.Command{Use: "sheet"}
+	read := &cobra.Command{Use: "read <range>"}
+	sheet.AddCommand(read)
+	root.AddCommand(sheet)
+	if got := telemetryCommandPath(root); got != "dws" {
+		t.Fatalf("root command path = %q, want dws", got)
+	}
+	if got := telemetryCommandPath(read); got != "sheet read" {
+		t.Fatalf("leaf command path = %q, want sheet read", got)
+	}
+	root.PersistentFlags().String("profile", "", "")
+	read.Aliases = []string{"get"}
+	if got := telemetryCommandPathForArgs(root, []string{"--profile", "corp-a", "sheet", "get", "A1:B2"}); got != "sheet read" {
+		t.Fatalf("pre-execution command path = %q, want sheet read", got)
+	}
+	if got := telemetryCommandPathForArgs(root, []string{"missing"}); got != "dws" {
+		t.Fatalf("unknown pre-execution command path = %q, want dws", got)
 	}
 }
 

internal/app/root_help.go

@@ -2,6 +2,7 @@ package app
 
 import (
 	"fmt"
+	"io"
 	"strings"
 	"text/tabwriter"
 
@@ -13,6 +14,12 @@ import (
 	"github.com/spf13/pflag"
 )
 
+// feedbackFormURL points at the DingTalk Notable form collecting dws CLI
+// user-experience feedback. The source parameter tags submissions that
+// originated from the CLI help output so they can be told apart from
+// responses arriving through other channels.
+const feedbackFormURL = "https://alidocs.dingtalk.com/notable/share/form/v01eLbnj1bw1ELb0laN_dv19yqvsgs3oebp3pcjys_1qX0QQ0?source=dws-cli"
+
 func configureRootHelp(root *cobra.Command) {
 	if root == nil {
 		return
@@ -101,6 +108,29 @@ func renderRootHelp(root *cobra.Command) {
 		_, _ = fmt.Fprintln(w)
 		_, _ = fmt.Fprintln(w, tui.Dim(long))
 	}
+
+	// Keep the feedback entry last: everything above it is operational guidance
+	// an agent acts on, while the survey is addressed to human readers who
+	// scroll to the end.
+	_, _ = fmt.Fprintln(w)
+	renderRootFeedback(w)
+}
+
+// renderRootFeedback prints the user-experience survey entry. The URL occupies
+// its own line and is never wrapped or padded through a tabwriter: it is longer
+// than the help rule width, and breaking it would stop terminals from
+// recognizing it as a clickable hyperlink. Soft wrapping performed by the
+// terminal itself keeps the link intact.
+//
+// The label is intentionally not routed through i18n. Everything surrounding it
+// in this listing — service descriptions, utility descriptions, global flag
+// usage — is hardcoded Chinese, so translating this one line would render it in
+// English on any host whose LANG is not zh_*, leaving a single English line
+// inside an otherwise Chinese screen.
+func renderRootFeedback(w io.Writer) {
+	_, _ = fmt.Fprintln(w, tui.Section("Feedback:"))
+	_, _ = fmt.Fprintf(w, "  %s %s\n", tui.Bullet(), tui.Dim("使用体验反馈问卷（1 分钟）"))
+	_, _ = fmt.Fprintf(w, "    %s\n", tui.Cyan(feedbackFormURL))
 }
 
 func renderRootGlobalFlags(root *cobra.Command) {

internal/app/root_help_test.go

@@ -53,6 +53,49 @@ func TestRootHelpHidesCompatibilityOnlyCommands(t *testing.T) {
 	}
 }
 
+func TestRootHelpShowsFeedbackEntry(t *testing.T) {
+	cmd := NewRootCommand()
+	var out bytes.Buffer
+	cmd.SetOut(&out)
+	cmd.SetErr(&out)
+	cmd.SetArgs([]string{"--help"})
+	if err := cmd.Execute(); err != nil {
+		t.Fatalf("root help: %v\n%s", err, out.String())
+	}
+	help := out.String()
+	// The label stays Chinese regardless of the host locale: the rest of this
+	// listing is hardcoded Chinese, so a translated label would show up as a
+	// lone English line on any host whose LANG is not zh_*.
+	for _, want := range []string{"Feedback:", "使用体验反馈问卷", feedbackFormURL} {
+		if !strings.Contains(help, want) {
+			t.Fatalf("root help missing %q:\n%s", want, help)
+		}
+	}
+	// The form URL is longer than the help rule width; it must stay on a
+	// single unbroken line so terminals keep recognizing it as a hyperlink.
+	if !strings.Contains(help, "\n    "+feedbackFormURL+"\n") {
+		t.Fatalf("feedback URL must occupy one unwrapped line:\n%s", help)
+	}
+}
+
+// The feedback entry is deliberately root-only: this CLI is driven mostly by
+// AI agents, and repeating a survey link in every subcommand help would be
+// pure context noise. Guard the boundary so a future refactor cannot move the
+// rendering into the shared subcommand help path unnoticed.
+func TestSubcommandHelpOmitsFeedbackEntry(t *testing.T) {
+	cmd := NewRootCommand()
+	var out bytes.Buffer
+	cmd.SetOut(&out)
+	cmd.SetErr(&out)
+	cmd.SetArgs([]string{"chat", "--help"})
+	if err := cmd.Execute(); err != nil {
+		t.Fatalf("chat help: %v\n%s", err, out.String())
+	}
+	if help := out.String(); strings.Contains(help, feedbackFormURL) {
+		t.Fatalf("subcommand help must not carry the feedback URL:\n%s", help)
+	}
+}
+
 func TestCalendarEventCreateHelpKeepsRoomsStringMetavar(t *testing.T) {
 	cmd := NewRootCommand()
 	var out bytes.Buffer
@@ -221,7 +264,7 @@ func TestRootChatMediaUploadWithoutAppCredentialsReturnsMigrationValidation(t *t
 	}
 
 	got := output.String() + "\n" + err.Error()
-	for _, want := range []string{"已下线", "chat message send --msg-type file --file-path"} {
+	for _, want := range []string{"已下线", "chat message send --msg-type file --file"} {
 		if !strings.Contains(got, want) {
 			t.Fatalf("chat media upload migration output missing %q:\n%s", want, got)
 		}
@@ -396,7 +439,7 @@ func TestChatFileUploadDownlinedButMessageFileSendStays(t *testing.T) {
 		t.Fatalf("chat file upload error = nil, want downline error\n%s", got)
 	}
 	got = got + "\n" + err.Error()
-	for _, want := range []string{"已下线", "upload_conversation_file_by_url", "chat message send --msg-type file --file-path"} {
+	for _, want := range []string{"已下线", "upload_conversation_file_by_url", "chat message send --msg-type file --file"} {
 		if !strings.Contains(got, want) {
 			t.Fatalf("chat file upload output missing %q:\n%s", want, got)
 		}
@@ -419,6 +462,53 @@ func TestCalendarEventListDryRunPreviewsOnly(t *testing.T) {
 	}
 }
 
+func TestCalendarEventShareInfoDryRunPreviewsOnly(t *testing.T) {
+	got, err := executeRootCaptureStdout(t, []string{
+		"--dry-run", "calendar", "event", "share-info",
+		"--id", "EVT_001",
+		"--language", "zh-CN",
+		"--calendar-id", "primary",
+	})
+	if err != nil {
+		t.Fatalf("calendar event share-info --dry-run error = %v\n%s", err, got)
+	}
+	for _, want := range []string{"get_event_share_info", "eventId", "EVT_001", "zh-CN", "primary"} {
+		if !strings.Contains(got, want) {
+			t.Fatalf("calendar event share-info dry-run output missing %q:\n%s", want, got)
+		}
+	}
+}
+
+func TestCalendarEventShareInfoRequiresEventID(t *testing.T) {
+	got, err := executeRootCaptureStdout(t, []string{
+		"--dry-run", "calendar", "event", "share-info",
+	})
+	if err == nil {
+		t.Fatalf("calendar event share-info without --id: expected error, got nil\n%s", got)
+	}
+	if strings.Contains(got, "\"executed\": true") {
+		t.Fatalf("share-info without --id must not execute:\n%s", got)
+	}
+}
+
+func TestCalendarEventShareInfoOmitsOptionalArgs(t *testing.T) {
+	got, err := executeRootCaptureStdout(t, []string{
+		"--dry-run", "calendar", "event", "share-info",
+		"--id", "EVT_001",
+	})
+	if err != nil {
+		t.Fatalf("calendar event share-info --dry-run with only --id error = %v\n%s", err, got)
+	}
+	if !strings.Contains(got, "\"eventId\"") {
+		t.Fatalf("calendar event share-info dry-run output missing eventId:\n%s", got)
+	}
+	for _, unwanted := range []string{"\"calendarId\"", "\"language\""} {
+		if strings.Contains(got, unwanted) {
+			t.Fatalf("calendar event share-info dry-run with only --id should not contain %q:\n%s", unwanted, got)
+		}
+	}
+}
+
 func TestRootKeepsSVIPChatCompatibilityFlags(t *testing.T) {
 	root := NewRootCommand()
 

internal/app/root_lifecycle_regression_test.go

@@ -181,7 +181,7 @@ func TestPublicRootDirectExecuteClosesSinkOnHandlerError(t *testing.T) {
 	}
 }
 
-func TestExecutePanicAfterEmissionPreservesSingleResultAndExitCode(t *testing.T) {
+func TestCrossPlatformCoverageExecutePanicAfterEmissionPreservesSingleResultAndExitCode(t *testing.T) {
 	oldNormalize := rootNormalizeProcessProfileArgs
 	oldExecute := rootExecuteCommand
 	oldNewRoot := rootNewRootCommandWithEngine

internal/app/root_output_atomic_test.go

@@ -230,13 +230,20 @@ func TestOutputSinkUnifiedPublicationFailureFailsAndLeavesNoFinalFile(t *testing
 	assertNoOutputTemps(t, target)
 }
 
-func TestExecuteUnifiedPublicationFailureEmitsFailureOnOriginalStdout(t *testing.T) {
+func TestCrossPlatformCoverageExecuteUnifiedPublicationFailureEmitsFailureOnOriginalStdout(t *testing.T) {
 	testseam.Protect(t, &os.Args)
 	dir := t.TempDir()
-	target := filepath.Join(dir, "result.json")
+	t.Chdir(dir)
+	// Keep argv portable: an absolute Windows path contains a volume colon,
+	// which the CLI intentionally rejects as unsafe user-supplied output.
+	target := "result.json"
 	if err := os.WriteFile(target, []byte("original"), 0o640); err != nil {
 		t.Fatal(err)
 	}
+	originalInfo, err := os.Stat(target)
+	if err != nil {
+		t.Fatal(err)
+	}
 	os.Args = []string{"dws", "atomic-output-unified-publication", "--output", target, "--format", "json"}
 	testseam.Swap(t, &rootRenameFile, func(string, string) error { return errors.New("rename failed") })
 	testseam.Swap(t, &rootNormalizeProcessProfileArgs, func() func() { return func() {} })
@@ -277,7 +284,7 @@ func TestExecuteUnifiedPublicationFailureEmitsFailureOnOriginalStdout(t *testing
 	if got := bytes.Count(stdout.Bytes(), []byte(`"outcome": "success"`)); got != 0 {
 		t.Fatalf("rolled-back success leaked to stdout: %s", stdout.String())
 	}
-	assertOutputFile(t, target, "original", 0o640)
+	assertOutputFile(t, target, "original", originalInfo.Mode().Perm())
 	assertNoOutputTemps(t, target)
 }
 

internal/app/root_unified_output_test.go

@@ -27,7 +27,7 @@ import (
 	"github.com/spf13/cobra"
 )
 
-func TestExecuteEmitsStoredUnifiedResultAtSingleRootExit(t *testing.T) {
+func TestCrossPlatformCoverageExecuteEmitsStoredUnifiedResultAtSingleRootExit(t *testing.T) {
 	oldNormalize := rootNormalizeProcessProfileArgs
 	oldExecute := rootExecuteCommand
 	oldNewRoot := rootNewRootCommandWithEngine

internal/app/schema_calendar_shortcut_contract_test.go

@@ -0,0 +1,49 @@
+// Copyright 2026 Alibaba Group
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+
+package app
+
+import "testing"
+
+func TestCrossPlatformCoverageCalendarAgendaFinalSchemaPreservesCompositeProperties(t *testing.T) {
+	snapshot := fullSchemaSnapshotForTest(t)
+	tool := snapshot.Tools["calendar.shortcut_agenda"]
+	if tool == nil {
+		t.Fatal("calendar.shortcut_agenda is missing from final Schema")
+	}
+	parameters := schemaContractMap(tool["parameters"])
+	for flag, want := range map[string]string{
+		"start": "start",
+		"end":   "end",
+	} {
+		parameter := parameters[flag]
+		if parameter == nil {
+			t.Fatalf("calendar.shortcut_agenda --%s is missing from final Schema", flag)
+		}
+		if got := schemaContractString(parameter["property"]); got != want {
+			t.Errorf("calendar.shortcut_agenda --%s property=%q, want %q", flag, got, want)
+		}
+	}
+	if got := schemaContractString(tool["interface_mode"]); got != "composite" {
+		t.Fatalf("calendar.shortcut_agenda interface_mode=%q, want composite", got)
+	}
+	result := schemaContractMap(tool["result"])
+	dataSchema := schemaContractMap(result["data_schema"])
+	properties := schemaContractMap(dataSchema["properties"])
+	for _, field := range []string{"hasMore", "nextCursor"} {
+		if _, exists := properties[field]; exists {
+			t.Fatalf("calendar.shortcut_agenda Result data_schema leaked pagination field %q", field)
+		}
+	}
+	if properties["complete"] == nil {
+		t.Fatal("calendar.shortcut_agenda Result data_schema is missing complete")
+	}
+	pagination, ok := tool["pagination"].(map[string]any)
+	if !ok {
+		t.Fatalf("calendar.shortcut_agenda pagination=%T, want object", tool["pagination"])
+	}
+	if got := schemaContractString(pagination["meta_path"]); got != "meta.pagination" {
+		t.Fatalf("calendar.shortcut_agenda pagination meta_path=%q, want meta.pagination", got)
+	}
+}

internal/app/schema_contract_test.go

@@ -400,6 +400,7 @@ func schemaContractPayloadForBoundCanonicals(t *testing.T, root *cobra.Command,
 func TestChatSchemaSeparatesSendAndReply(t *testing.T) {
 	snapshot := schemaContractPayloadForBoundCanonicals(t, NewRootCommand(),
 		"chat.send_personal_message",
+		"chat.send_robot_message",
 		"chat.reply_personal_message",
 	)
 
@@ -418,6 +419,19 @@ func TestChatSchemaSeparatesSendAndReply(t *testing.T) {
 	if _, exists := snapshot.Tools["chat.upload_conversation_file"]; exists {
 		t.Fatal("downlined chat file upload must not be advertised in Schema")
 	}
+
+	botReply := snapshot.Tools["chat.send_robot_message"]
+	botParams := schemaContractMap(botReply["parameters"])
+	if got := schemaContractString(botParams["reply"]["property"]); got != "referenceOpenMessageId" {
+		t.Fatalf("bot --reply property = %q", got)
+	}
+	if got := schemaContractString(botParams["ref-sender"]["property"]); got != "srcMsgSendOpenDingTalkId" {
+		t.Fatalf("bot --ref-sender property = %q", got)
+	}
+	if got, ok := botParams["title"]["required"].(bool); !ok || got {
+		t.Fatalf("bot --title required = %#v, want false for conditional Markdown input", botParams["title"]["required"])
+	}
+	assertSchemaContractConstraintGroup(t, botReply, "require_together", []string{"reply", "ref-sender"})
 }
 
 func TestCalendarAttendeeDeleteSchemaMatchesRuntimeGate(t *testing.T) {

internal/app/schema_oa_attachment_contract_test.go

@@ -0,0 +1,166 @@
+// Copyright 2026 Alibaba Group
+// Licensed under the Apache License, Version 2.0 (the "License");
+
+package app
+
+import (
+	"bytes"
+	"encoding/json"
+	"reflect"
+	"strings"
+	"testing"
+)
+
+func TestCrossPlatformCoverageOAAttachmentDeliveredSchemaMatchesExecutableHelp(t *testing.T) {
+	tests := []struct {
+		cliPath        string
+		canonical      string
+		rpc            string
+		description    string
+		effect         string
+		resultType     string
+		resultFields   map[string]string
+		sensitivePaths []string
+	}{
+		{
+			cliPath:     "oa approval attachment download-url",
+			canonical:   "oa.get_attachment_download_url",
+			rpc:         "get_attachment_download_url",
+			description: "获取审批附件下载授权并生成临时下载链接",
+			effect:      "read",
+			resultType:  "object",
+			resultFields: map[string]string{
+				"spaceId": "integer", "agentId": "integer", "downloadUri": "string",
+				"class": "string", "fileId": "string",
+			},
+			sensitivePaths: []string{"downloadUri"},
+		},
+		{
+			cliPath:     "oa approval attachment authorize-download",
+			canonical:   "oa.auth_download_file",
+			rpc:         "auth_download_file",
+			description: "批量授权当前用户下载指定的审批钉盘文件",
+			effect:      "write",
+			resultType:  "boolean",
+		},
+		{
+			cliPath:     "oa approval attachment authorize-preview",
+			canonical:   "oa.auth_preview_attachment",
+			rpc:         "auth_preview_attachment",
+			description: "批量授权当前用户预览审批单中的附件",
+			effect:      "write",
+			resultType:  "object",
+			resultFields: map[string]string{
+				"spaceId": "integer", "agentId": "integer", "class": "string",
+			},
+		},
+	}
+
+	for _, test := range tests {
+		t.Run(test.rpc, func(t *testing.T) {
+			root := NewRootCommand()
+			command := exactCommandForTest(root, test.cliPath)
+			if command == nil {
+				t.Fatalf("executable command %q is missing", test.cliPath)
+			}
+
+			var stdout, stderr bytes.Buffer
+			root.SetOut(&stdout)
+			root.SetErr(&stderr)
+			root.SetArgs([]string{"schema", test.cliPath, "--format", "json"})
+			if err := root.Execute(); err != nil {
+				t.Fatalf("execute delivery schema leaf: %v; stderr=%s", err, stderr.String())
+			}
+
+			var tool map[string]any
+			if err := json.Unmarshal(stdout.Bytes(), &tool); err != nil {
+				t.Fatalf("decode delivery schema leaf: %v", err)
+			}
+			if got := schemaContractString(tool["canonical_path"]); got != test.canonical {
+				t.Fatalf("canonical_path = %q, want %q", got, test.canonical)
+			}
+			if got := schemaContractString(tool["primary_cli_path"]); got != test.cliPath {
+				t.Fatalf("primary_cli_path = %q, want %q", got, test.cliPath)
+			}
+			if got := schemaContractString(tool["description"]); !strings.HasPrefix(got, test.description) {
+				t.Fatalf("description = %q, want prefix %q", got, test.description)
+			}
+			if got := schemaContractString(tool["interface_mode"]); got != "mcp" {
+				t.Fatalf("interface_mode = %q, want mcp", got)
+			}
+			if got := schemaContractString(tool["availability"]); got != "available" {
+				t.Fatalf("availability = %q, want available", got)
+			}
+			interfaceRef := schemaInterfaceObject(tool["interface_ref"])
+			if got := schemaContractString(interfaceRef["product_id"]); got != "oa" {
+				t.Fatalf("interface_ref.product_id = %q, want oa", got)
+			}
+			if got := schemaContractString(interfaceRef["rpc_name"]); got != test.rpc {
+				t.Fatalf("interface_ref.rpc_name = %q, want %q", got, test.rpc)
+			}
+			if got := schemaContractString(tool["effect"]); got != test.effect {
+				t.Fatalf("effect = %q, want %q", got, test.effect)
+			}
+			if got := schemaContractString(tool["risk"]); got != "low" {
+				t.Fatalf("risk = %q, want low", got)
+			}
+			if got := schemaContractString(tool["confirmation"]); got != "not_required" {
+				t.Fatalf("confirmation = %q, want not_required", got)
+			}
+			if got := schemaContractString(tool["idempotency"]); got != "idempotent" {
+				t.Fatalf("idempotency = %q, want idempotent", got)
+			}
+			fullResult := oaAttachmentResultContract(t, tool, test.resultType, test.resultFields, test.sensitivePaths)
+
+			stdout.Reset()
+			stderr.Reset()
+			root.SetArgs([]string{"schema", test.cliPath, "--compact", "--format", "json"})
+			if err := root.Execute(); err != nil {
+				t.Fatalf("execute compact delivery schema leaf: %v; stderr=%s", err, stderr.String())
+			}
+			var compactTool map[string]any
+			if err := json.Unmarshal(stdout.Bytes(), &compactTool); err != nil {
+				t.Fatalf("decode compact delivery schema leaf: %v", err)
+			}
+			compactResult, ok := compactTool["result"].(map[string]any)
+			if !ok {
+				t.Fatalf("compact result = %#v, want object", compactTool["result"])
+			}
+			if !reflect.DeepEqual(compactResult, fullResult) {
+				t.Fatalf("compact/full result projection differs\ncompact: %#v\nfull: %#v", compactResult, fullResult)
+			}
+			if problem := schemaHelpFlagCompletenessProblem(test.canonical, test.cliPath, command, tool); problem != "" {
+				t.Fatal(problem)
+			}
+		})
+	}
+}
+
+func oaAttachmentResultContract(t *testing.T, tool map[string]any, resultType string, fields map[string]string, sensitivePaths []string) map[string]any {
+	t.Helper()
+	result, ok := tool["result"].(map[string]any)
+	if !ok {
+		t.Fatalf("full result = %#v, want object", tool["result"])
+	}
+	if got, want := schemaContractStringSlice(result["outcomes"]), []string{"success", "failure"}; !reflect.DeepEqual(got, want) {
+		t.Fatalf("result.outcomes = %#v, want %#v", got, want)
+	}
+	if got := schemaContractStringSlice(result["sensitive_paths"]); !reflect.DeepEqual(got, sensitivePaths) {
+		t.Fatalf("result.sensitive_paths = %#v, want %#v", got, sensitivePaths)
+	}
+	dataSchema, ok := result["data_schema"].(map[string]any)
+	if !ok || schemaContractString(dataSchema["type"]) != resultType {
+		t.Fatalf("result.data_schema = %#v, want type %q", result["data_schema"], resultType)
+	}
+	properties, _ := dataSchema["properties"].(map[string]any)
+	if len(properties) != len(fields) {
+		t.Fatalf("result.data_schema.properties = %#v, want fields %#v", properties, fields)
+	}
+	for name, fieldType := range fields {
+		property, ok := properties[name].(map[string]any)
+		if !ok || schemaContractString(property["type"]) != fieldType {
+			t.Fatalf("result.data_schema.properties.%s = %#v, want type %q", name, properties[name], fieldType)
+		}
+	}
+	return result
+}

internal/app/schema_shortcut_contract_test.go

@@ -16,12 +16,12 @@ import (
 )
 
 const (
-	publicShortcutCount = 399
+	publicShortcutCount = 435
 	// schemaPublishedShortcutCount counts every delivered *.shortcut_* tool,
 	// including the hidden historical minutes.shortcut_minutes_search contract.
-	schemaPublishedShortcutCount = 401
+	schemaPublishedShortcutCount = 438
 	// publiclyDeliveredShortcutCount is the public-catalog subset of that surface.
-	publiclyDeliveredShortcutCount = 399
+	publiclyDeliveredShortcutCount = 435
 )
 
 func TestDeliverySchemaCoversOrExactlyExcludesEveryPublicShortcutContract(t *testing.T) {
@@ -140,6 +140,97 @@ func TestDeliveryShortcutProgressiveQueriesReturnCompleteContracts(t *testing.T)
 	assertChatCatalogCompleteLeafContracts(t)
 }
 
+func TestCrossPlatformCoverageAITableTableBootstrapPublishesResultContract(t *testing.T) {
+	leaf := executeShortcutSchemaQuery(t, "--cli-path", "aitable +table-bootstrap")
+	result, _ := leaf["result"].(map[string]any)
+	if got, want := schemaContractStringSlice(result["outcomes"]), []string{"success", "failure"}; !schemaContractJSONEqual(got, want) {
+		t.Fatalf("aitable +table-bootstrap outcomes = %#v, want %#v", got, want)
+	}
+	dataSchema, _ := result["data_schema"].(map[string]any)
+	properties := schemaContractMap(dataSchema["properties"])
+	status := properties["status"]
+	if got, want := schemaContractStringSlice(status["enum"]), []string{"success", "planned", "partial_success", "unknown"}; !schemaContractJSONEqual(got, want) {
+		t.Fatalf("aitable +table-bootstrap status enum = %#v, want %#v", got, want)
+	}
+	for _, property := range []string{"contractVersion", "operation", "executed", "retryable", "plan", "completedSteps", "verification", "checkpoint", "knownSideEffects", "result"} {
+		if properties[property] == nil {
+			t.Errorf("aitable +table-bootstrap final Result data_schema is missing %q", property)
+		}
+	}
+}
+
+func TestDeliveryWikiSpaceSearchDeclaresCompatibilityAdapter(t *testing.T) {
+	leaf := executeShortcutSchemaQuery(t, "--cli-path", "wiki +space-search")
+	if got := schemaContractString(leaf["interface_mode"]); got != "composite" {
+		t.Fatalf("wiki +space-search interface_mode = %q, want composite", got)
+	}
+	reason := schemaContractString(leaf["interface_reason"])
+	for _, fragment := range []string{"query/limit", "search_wikiSpaces.keyword/pageSize", "versioned Schema migration"} {
+		if !strings.Contains(reason, fragment) {
+			t.Fatalf("wiki +space-search interface_reason = %q, want fragment %q", reason, fragment)
+		}
+	}
+	parameters := schemaContractMap(leaf["parameters"])
+	for name, want := range map[string]string{"query": "query", "limit": "limit"} {
+		parameter := parameters[name]
+		if parameter == nil {
+			t.Fatalf("wiki +space-search missing --%s parameter: %#v", name, parameters)
+		}
+		if got := schemaContractString(parameter["property"]); got != want {
+			t.Fatalf("wiki +space-search --%s property = %q, want compatibility value %q", name, got, want)
+		}
+	}
+}
+
+func TestAllShortcutsWikiSchemaExamplesIncludeRequiredParameters(t *testing.T) {
+	tools := deliverySchemaAllToolsForHelpFlagTest(t, NewRootCommand())
+	checked := 0
+	for _, declared := range shortcut.All() {
+		if declared.Service != "wiki" || declared.UserDefined || !shortcut.InPublicCatalog(declared.Service, declared.Command) {
+			continue
+		}
+		checked++
+		canonical := shortcutSchemaCanonical(declared)
+		tool := tools[canonical]
+		if tool == nil {
+			t.Fatalf("delivery schema --all is missing %s", canonical)
+		}
+		examples := schemaContractStringSlice(tool["examples"])
+		if len(examples) == 0 {
+			t.Fatalf("%s has no delivered examples", canonical)
+		}
+		for _, example := range examples {
+			argv, err := cli.ParseAgentExampleArgv(example)
+			if err != nil {
+				t.Fatalf("%s example %q is not valid argv: %v", canonical, example, err)
+			}
+			for _, flag := range declared.Flags {
+				if !flag.Required {
+					continue
+				}
+				names := append([]string{flag.Name}, flag.Aliases...)
+				if !schemaExampleHasLongFlag(argv, names...) {
+					t.Errorf("%s example %q is missing required --%s", canonical, example, flag.Name)
+				}
+			}
+		}
+	}
+	if checked != 20 {
+		t.Fatalf("checked Wiki shortcut examples = %d, want 20", checked)
+	}
+}
+
+func schemaExampleHasLongFlag(argv []string, names ...string) bool {
+	for _, argument := range argv {
+		for _, name := range names {
+			if argument == "--"+name || strings.HasPrefix(argument, "--"+name+"=") {
+				return true
+			}
+		}
+	}
+	return false
+}
+
 func assertSchemaSummarySafety(
 	t testing.TB,
 	summaries map[string]map[string]any,

internal/app/signal_test.go

@@ -103,7 +103,7 @@ func installSignalExecuteSeams(t *testing.T, unified bool, stdout, stderr io.Wri
 	})
 }
 
-func TestExecuteSignalEmitsOneTypedUnifiedFailure(t *testing.T) {
+func TestCrossPlatformCoverageExecuteSignalEmitsOneTypedUnifiedFailure(t *testing.T) {
 	for _, tc := range []struct {
 		name    string
 		signal  syscall.Signal
@@ -197,7 +197,7 @@ func TestSignalAfterFailedEmissionAttemptPreservesPublicationExitCode(t *testing
 	}
 }
 
-func TestSignalBeforeEmissionAttemptPreservesPublishedOutcome(t *testing.T) {
+func TestCrossPlatformCoverageSignalBeforeEmissionAttemptPreservesPublishedOutcome(t *testing.T) {
 	var stdout bytes.Buffer
 	installSignalExecuteSeams(t, true, &stdout, io.Discard)
 	testseam.Swap(t, &rootExecuteCommand, func(cmd *cobra.Command) (*cobra.Command, error) {
@@ -229,7 +229,7 @@ func TestSignalBeforeEmissionAttemptPreservesPublishedOutcome(t *testing.T) {
 	}
 }
 
-func TestSignalAfterCompletedPrimaryPreservesEstablishedOutcome(t *testing.T) {
+func TestCrossPlatformCoverageSignalAfterCompletedPrimaryPreservesEstablishedOutcome(t *testing.T) {
 	var stdout bytes.Buffer
 	installSignalExecuteSeams(t, true, &stdout, io.Discard)
 	testseam.Swap(t, &rootExecuteCommand, func(cmd *cobra.Command) (*cobra.Command, error) {

internal/app/telemetry_error.go

@@ -0,0 +1,143 @@
+// Copyright 2026 Alibaba Group
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+package app
+
+import (
+	stderrors "errors"
+	"fmt"
+	"regexp"
+	"strings"
+	"unicode"
+
+	apperrors "github.com/DingTalk-Real-AI/dingtalk-workspace-cli/internal/errors"
+	"github.com/DingTalk-Real-AI/dingtalk-workspace-cli/internal/output"
+)
+
+const maxTelemetryErrorRunes = 200
+
+var (
+	telemetryUnknownFlagPattern  = regexp.MustCompile(`(?i)unknown flag:\s*(--[a-z0-9][a-z0-9-]*)`)
+	telemetryAuthPattern         = regexp.MustCompile(`(?i)\b(?:bearer|basic)\s+[a-z0-9._~+/=-]+`)
+	telemetrySensitiveFlag       = regexp.MustCompile(`(?i)(--(?:access-token|refresh-token|token|client-secret|client-id|password|api-key|authorization|cookie|credential|secret))(?:=|\s+)\S+`)
+	telemetrySensitiveValue      = regexp.MustCompile(`(?i)\b(authorization|client[-_]?secret|client[-_]?id|access[-_]?token|refresh[-_]?token|api[-_]?key|password|cookie|credential|secret|token)\b\s*[:=]\s*[^\s,;]+`)
+	telemetryURLPattern          = regexp.MustCompile(`(?i)\b(?:https?|wss?)://[^\s]+`)
+	telemetryJSONPattern         = regexp.MustCompile(`(?s)[\[{].*[\]}]`)
+	telemetryUnixPathPattern     = regexp.MustCompile(`(^|[\s=:])(?:~/|/)[^\s]+`)
+	telemetryWindowsPathPattern  = regexp.MustCompile(`(?i)(^|[\s=])[a-z]:[\\/][^\s]+`)
+	telemetryRelativePathPattern = regexp.MustCompile(`(^|[\s=:])\.\.?/[^\s]+`)
+	telemetryEmailPattern        = regexp.MustCompile(`(?i)\b[a-z0-9._%+-]+@[a-z0-9.-]+\.[a-z]{2,}\b`)
+	telemetryPhonePattern        = regexp.MustCompile(`\b\+?\d[\d -]{7,}\d\b`)
+	telemetryOpaqueTokenPattern  = regexp.MustCompile(`\b[a-zA-Z0-9_-]{16,}\b`)
+)
+
+func telemetryErrorSummary(err error) string {
+	if err == nil {
+		return ""
+	}
+	var patError *apperrors.PATError
+	if stderrors.As(err, &patError) {
+		return "permission error"
+	}
+	var rawError apperrors.RawStderrError
+	if stderrors.As(err, &rawError) {
+		return "raw stderr error"
+	}
+	if isUnknownCommandError(err) {
+		return "unknown command"
+	}
+	if match := telemetryUnknownFlagPattern.FindStringSubmatch(err.Error()); len(match) == 2 {
+		return "unknown flag: " + match[1]
+	}
+	return sanitizeTelemetryErrorText(err.Error())
+}
+
+func telemetryPanicMessages(value any) (display, summary string) {
+	return fmt.Sprintf("internal panic: %v", value), "internal panic"
+}
+
+func sanitizeTelemetryErrorText(message string) string {
+	message = output.SanitizeForTerminal(message)
+	message = telemetryAuthPattern.ReplaceAllString(message, "<credential>")
+	message = telemetrySensitiveFlag.ReplaceAllString(message, "$1=<redacted>")
+	message = telemetrySensitiveValue.ReplaceAllString(message, "$1=<redacted>")
+	message = telemetryURLPattern.ReplaceAllString(message, "<url>")
+	message = telemetryJSONPattern.ReplaceAllString(message, "<payload>")
+	message = redactTelemetryQuotedText(message)
+	message = telemetryUnixPathPattern.ReplaceAllString(message, "$1<path>")
+	message = telemetryWindowsPathPattern.ReplaceAllString(message, "$1<path>")
+	message = telemetryRelativePathPattern.ReplaceAllString(message, "$1<path>")
+	message = telemetryEmailPattern.ReplaceAllString(message, "<email>")
+	message = telemetryPhonePattern.ReplaceAllString(message, "<phone>")
+	message = telemetryOpaqueTokenPattern.ReplaceAllStringFunc(message, func(value string) string {
+		var hasLetter, hasDigit bool
+		for _, r := range value {
+			hasLetter = hasLetter || unicode.IsLetter(r)
+			hasDigit = hasDigit || unicode.IsDigit(r)
+		}
+		if hasLetter && hasDigit {
+			return "<id>"
+		}
+		return value
+	})
+	message = strings.Join(strings.Fields(message), " ")
+	return truncateTelemetryText(message, maxTelemetryErrorRunes)
+}
+
+func redactTelemetryQuotedText(message string) string {
+	var result strings.Builder
+	runes := []rune(message)
+	for index := 0; index < len(runes); {
+		quote := runes[index]
+		if quote != '\'' && quote != '"' && quote != '`' {
+			result.WriteRune(quote)
+			index++
+			continue
+		}
+		result.WriteRune(quote)
+		result.WriteString("<redacted>")
+		index++
+		escaped := false
+		for index < len(runes) {
+			current := runes[index]
+			index++
+			if escaped {
+				escaped = false
+				continue
+			}
+			if current == '\\' && quote != '`' {
+				escaped = true
+				continue
+			}
+			if current == quote {
+				result.WriteRune(quote)
+				break
+			}
+		}
+	}
+	return result.String()
+}
+
+func truncateTelemetryText(message string, maxRunes int) string {
+	if maxRunes <= 0 {
+		return ""
+	}
+	runes := []rune(message)
+	if len(runes) <= maxRunes {
+		return message
+	}
+	if maxRunes <= 3 {
+		return string(runes[:maxRunes])
+	}
+	return string(runes[:maxRunes-3]) + "..."
+}

internal/app/telemetry_error_test.go

@@ -0,0 +1,99 @@
+// Copyright 2026 Alibaba Group
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+package app
+
+import (
+	"errors"
+	"strings"
+	"testing"
+
+	apperrors "github.com/DingTalk-Real-AI/dingtalk-workspace-cli/internal/errors"
+)
+
+type telemetryRawError string
+
+func (e telemetryRawError) Error() string     { return string(e) }
+func (e telemetryRawError) RawStderr() string { return string(e) }
+
+func TestCrossPlatformCoverageTelemetryErrorSummaryFixedFamilies(t *testing.T) {
+	for _, tc := range []struct {
+		name string
+		err  error
+		want string
+	}{
+		{name: "nil", want: ""},
+		{name: "PAT", err: &apperrors.PATError{RawJSON: `{"token":"secret"}`}, want: "permission error"},
+		{name: "raw stderr", err: telemetryRawError("raw secret"), want: "raw stderr error"},
+		{name: "unknown command", err: errors.New(`unknown command "secret-value" for "dws"`), want: "unknown command"},
+		{name: "unknown flag", err: errors.New("unknown flag: --token=secret-value"), want: "unknown flag: --token"},
+	} {
+		t.Run(tc.name, func(t *testing.T) {
+			if got := telemetryErrorSummary(tc.err); got != tc.want {
+				t.Fatalf("telemetryErrorSummary() = %q, want %q", got, tc.want)
+			}
+		})
+	}
+}
+
+func TestCrossPlatformCoverageSanitizeTelemetryErrorText(t *testing.T) {
+	message := "\x1b[31mfailed\x1b[0m " +
+		"--client-secret very-secret " +
+		"--access-token access-secret " +
+		"Authorization: Bearer abcdefghijklmnop1234 " +
+		"url=https://example.test/path?token=secret " +
+		`body={"access_token":"secret"} ` +
+		`user="Alice" email=alice@example.test phone=13800138000 ` +
+		"path=/Users/alice/private.txt relative=./private/secrets.txt id=abcDEF1234567890XYZ"
+	got := sanitizeTelemetryErrorText(message)
+	for _, secret := range []string{
+		"very-secret", "access-secret", "abcdefghijklmnop1234", "example.test", "access_token",
+		"Alice", "alice@example.test", "13800138000", "/Users/alice", "abcDEF1234567890XYZ", "\x1b",
+		"./private/secrets.txt",
+	} {
+		if strings.Contains(got, secret) {
+			t.Fatalf("sanitized telemetry error leaked %q: %q", secret, got)
+		}
+	}
+	for _, marker := range []string{"failed", "<redacted>", "<url>", "<payload>", "<path>", "<id>"} {
+		if !strings.Contains(got, marker) {
+			t.Fatalf("sanitized telemetry error missing %q: %q", marker, got)
+		}
+	}
+}
+
+func TestCrossPlatformCoverageTelemetryErrorTruncationAndPanic(t *testing.T) {
+	message := strings.Repeat("错", maxTelemetryErrorRunes+1)
+	got := sanitizeTelemetryErrorText(message)
+	if len([]rune(got)) != maxTelemetryErrorRunes || !strings.HasSuffix(got, "...") {
+		t.Fatalf("truncated telemetry error rune length = %d suffix = %q", len([]rune(got)), got[len(got)-3:])
+	}
+	display, summary := telemetryPanicMessages("token-secret")
+	if display != "internal panic: token-secret" || summary != "internal panic" || strings.Contains(summary, "token-secret") {
+		t.Fatalf("panic messages = display %q summary %q", display, summary)
+	}
+	if got := truncateTelemetryText("value", 0); got != "" {
+		t.Fatalf("zero-limit truncation = %q", got)
+	}
+	if got := truncateTelemetryText("value", 3); got != "val" {
+		t.Fatalf("short-limit truncation = %q", got)
+	}
+}
+
+func TestCrossPlatformCoverageTelemetryErrorEscapesAndOpaqueWords(t *testing.T) {
+	const opaqueWord = "abcdefghijklmnop"
+	got := sanitizeTelemetryErrorText(`failed "quoted \"value" ` + opaqueWord)
+	if strings.Contains(got, "value") || !strings.Contains(got, opaqueWord) {
+		t.Fatalf("escaped quote sanitization = %q", got)
+	}
+}

internal/app/telemetry_identity.go

@@ -0,0 +1,86 @@
+// Copyright 2026 Alibaba Group
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+// http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+package app
+
+import (
+	"fmt"
+	"strings"
+
+	authpkg "github.com/DingTalk-Real-AI/dingtalk-workspace-cli/internal/auth"
+)
+
+// TelemetryIdentity is the privacy-reviewed subset of the local authentication
+// record that may be attached to a CLI execution event.
+type TelemetryIdentity struct {
+	UserID   string
+	UserName string
+	CorpID   string
+}
+
+var telemetryResolveProfileMetadata = authpkg.ResolveProfileMetadataReadOnly
+
+// ResolveTelemetryIdentity returns a pre-execution snapshot of the identity
+// selected by args. Multi-profile executions are attributed to the current
+// default profile. Resolution is deliberately best-effort: telemetry must not
+// refresh credentials or change command behavior when local auth data is
+// missing, invalid, or unreadable.
+func ResolveTelemetryIdentity(args []string) (identity TelemetryIdentity) {
+	defer func() {
+		if recover() != nil {
+			identity = TelemetryIdentity{}
+		}
+	}()
+
+	selector, specified, valid := preparseProfileSelection(args)
+	if specified && !valid {
+		return TelemetryIdentity{}
+	}
+	profile, err := resolveTelemetryProfileMetadata(defaultConfigDir(), selector)
+	if err != nil || profile == nil {
+		return TelemetryIdentity{}
+	}
+	return TelemetryIdentity{
+		UserID:   strings.TrimSpace(profile.UserID),
+		UserName: strings.TrimSpace(profile.UserName),
+		CorpID:   strings.TrimSpace(profile.CorpID),
+	}
+}
+
+func resolveTelemetryProfileMetadata(configDir, selector string) (*authpkg.ProfileMetadata, error) {
+	selector = strings.TrimSpace(selector)
+	if selector == "" || !strings.Contains(selector, ",") {
+		return telemetryResolveProfileMetadata(configDir, selector)
+	}
+
+	// A local profile name may itself contain a comma. Match the runtime
+	// resolver by trying the full selector before interpreting it as CSV.
+	if profile, err := telemetryResolveProfileMetadata(configDir, selector); err == nil && profile != nil {
+		return profile, nil
+	}
+
+	for _, part := range strings.Split(selector, ",") {
+		part = strings.TrimSpace(part)
+		if part == "" {
+			return nil, fmt.Errorf("--profile contains an empty profile selector: %q", selector)
+		}
+		profile, err := telemetryResolveProfileMetadata(configDir, part)
+		if err != nil {
+			return nil, err
+		}
+		if profile == nil {
+			return nil, fmt.Errorf("profile %q not found", part)
+		}
+	}
+	return telemetryResolveProfileMetadata(configDir, "")
+}

internal/app/telemetry_identity_test.go

@@ -0,0 +1,147 @@
+// Copyright 2026 Alibaba Group
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+package app
+
+import (
+	"errors"
+	"reflect"
+	"strings"
+	"testing"
+
+	authpkg "github.com/DingTalk-Real-AI/dingtalk-workspace-cli/internal/auth"
+	"github.com/DingTalk-Real-AI/dingtalk-workspace-cli/internal/testseam"
+)
+
+func TestCrossPlatformCoverageResolveTelemetryIdentityProfileSelection(t *testing.T) {
+	t.Setenv("DWS_CONFIG_DIR", "/telemetry-config")
+	profiles := map[string]*authpkg.ProfileMetadata{
+		"":           {UserID: " default-user ", UserName: " Default User ", CorpID: " default-corp "},
+		"corp-a":     {UserID: "user-a", UserName: "Alice", CorpID: "corp-a"},
+		"corp-b":     {UserID: "user-b", UserName: "Bob", CorpID: "corp-b"},
+		"alpha,beta": {UserID: "comma-user", UserName: "Comma User", CorpID: "comma-corp"},
+	}
+
+	for _, tc := range []struct {
+		name      string
+		args      []string
+		profiles  map[string]*authpkg.ProfileMetadata
+		wantCalls []string
+		want      TelemetryIdentity
+	}{
+		{name: "default profile", args: []string{"version"}, profiles: profiles, wantCalls: []string{""}, want: TelemetryIdentity{UserID: "default-user", UserName: "Default User", CorpID: "default-corp"}},
+		{name: "single profile", args: []string{"--profile", "corp-a", "version"}, profiles: profiles, wantCalls: []string{"corp-a"}, want: TelemetryIdentity{UserID: "user-a", UserName: "Alice", CorpID: "corp-a"}},
+		{name: "equals form after command", args: []string{"version", "--profile=corp-b"}, profiles: profiles, wantCalls: []string{"corp-b"}, want: TelemetryIdentity{UserID: "user-b", UserName: "Bob", CorpID: "corp-b"}},
+		{name: "last repeated profile", args: []string{"--profile", "corp-a", "version", "--profile=corp-b"}, profiles: profiles, wantCalls: []string{"corp-b"}, want: TelemetryIdentity{UserID: "user-b", UserName: "Bob", CorpID: "corp-b"}},
+		{name: "comma profile name", args: []string{"--profile", "alpha,beta", "version"}, profiles: profiles, wantCalls: []string{"alpha,beta"}, want: TelemetryIdentity{UserID: "comma-user", UserName: "Comma User", CorpID: "comma-corp"}},
+		{name: "multi profile uses default", args: []string{"--profile", "corp-a,corp-b", "version"}, profiles: profiles, wantCalls: []string{"corp-a,corp-b", "corp-a", "corp-b", ""}, want: TelemetryIdentity{UserID: "default-user", UserName: "Default User", CorpID: "default-corp"}},
+		{name: "unquoted multi profile", args: []string{"--profile", "corp-a,", "corp-b", "version"}, profiles: profiles, wantCalls: []string{"corp-a,corp-b", "corp-a", "corp-b", ""}, want: TelemetryIdentity{UserID: "default-user", UserName: "Default User", CorpID: "default-corp"}},
+	} {
+		t.Run(tc.name, func(t *testing.T) {
+			var calls []string
+			testseam.Swap(t, &telemetryResolveProfileMetadata, func(configDir, selector string) (*authpkg.ProfileMetadata, error) {
+				if configDir != "/telemetry-config" {
+					t.Fatalf("resolver config dir = %q", configDir)
+				}
+				calls = append(calls, selector)
+				profile := tc.profiles[selector]
+				if profile == nil {
+					return nil, errors.New("profile not found")
+				}
+				clone := *profile
+				return &clone, nil
+			})
+
+			if got := ResolveTelemetryIdentity(tc.args); got != tc.want {
+				t.Fatalf("ResolveTelemetryIdentity() = %#v, want %#v", got, tc.want)
+			}
+			if !reflect.DeepEqual(calls, tc.wantCalls) {
+				t.Fatalf("metadata selectors = %#v, want %#v", calls, tc.wantCalls)
+			}
+		})
+	}
+}
+
+func TestCrossPlatformCoverageResolveTelemetryIdentityRejectsMissingProfileValue(t *testing.T) {
+	for _, args := range [][]string{
+		{"version", "--profile"},
+		{"--profile=corp-a", "version", "--profile="},
+		{"--profile", "--debug", "version"},
+	} {
+		t.Run(strings.Join(args, "_"), func(t *testing.T) {
+			testseam.Swap(t, &telemetryResolveProfileMetadata, func(string, string) (*authpkg.ProfileMetadata, error) {
+				t.Fatal("invalid profile syntax attempted metadata resolution")
+				return nil, nil
+			})
+			if got := ResolveTelemetryIdentity(args); got != (TelemetryIdentity{}) {
+				t.Fatalf("invalid profile identity = %#v", got)
+			}
+		})
+	}
+}
+
+func TestCrossPlatformCoverageResolveTelemetryIdentityFailsClosed(t *testing.T) {
+	t.Setenv("DWS_CONFIG_DIR", "/telemetry-config")
+	fail := errors.New("metadata unavailable")
+
+	for _, tc := range []struct {
+		name    string
+		resolve func(string, string) (*authpkg.ProfileMetadata, error)
+	}{
+		{name: "read error", resolve: func(string, string) (*authpkg.ProfileMetadata, error) { return nil, fail }},
+		{name: "missing profile", resolve: func(string, string) (*authpkg.ProfileMetadata, error) { return nil, nil }},
+		{name: "empty fields", resolve: func(string, string) (*authpkg.ProfileMetadata, error) { return &authpkg.ProfileMetadata{}, nil }},
+		{name: "resolver panic", resolve: func(string, string) (*authpkg.ProfileMetadata, error) { panic("metadata failure") }},
+	} {
+		t.Run(tc.name, func(t *testing.T) {
+			testseam.Swap(t, &telemetryResolveProfileMetadata, tc.resolve)
+			if got := ResolveTelemetryIdentity(nil); got != (TelemetryIdentity{}) {
+				t.Fatalf("failed-closed identity = %#v", got)
+			}
+		})
+	}
+}
+
+func TestCrossPlatformCoverageResolveTelemetryIdentityRejectsInvalidMultiProfile(t *testing.T) {
+	t.Setenv("DWS_CONFIG_DIR", "/telemetry-config")
+	testseam.Swap(t, &telemetryResolveProfileMetadata, func(_ string, selector string) (*authpkg.ProfileMetadata, error) {
+		if selector == "corp-a" {
+			return &authpkg.ProfileMetadata{UserID: "user-a", CorpID: "corp-a"}, nil
+		}
+		return nil, errors.New("profile not found")
+	})
+	if got := ResolveTelemetryIdentity([]string{"--profile", "corp-a,missing", "version"}); got != (TelemetryIdentity{}) {
+		t.Fatalf("invalid multi-profile identity = %#v", got)
+	}
+}
+
+func TestCrossPlatformCoverageResolveTelemetryProfileMetadataRejectsMalformedMulti(t *testing.T) {
+	testseam.Swap(t, &telemetryResolveProfileMetadata, func(_ string, selector string) (*authpkg.ProfileMetadata, error) {
+		switch selector {
+		case "corp-a,,corp-b":
+			return nil, errors.New("not a literal profile")
+		case "corp-a":
+			return &authpkg.ProfileMetadata{UserID: "user-a"}, nil
+		case "missing":
+			return nil, nil
+		default:
+			return nil, errors.New("unexpected selector")
+		}
+	})
+	if _, err := resolveTelemetryProfileMetadata("/config", "corp-a,,corp-b"); err == nil || !strings.Contains(err.Error(), "empty profile selector") {
+		t.Fatalf("empty multi-profile selector error = %v", err)
+	}
+	if _, err := resolveTelemetryProfileMetadata("/config", "corp-a,missing"); err == nil || !strings.Contains(err.Error(), "not found") {
+		t.Fatalf("missing multi-profile selector error = %v", err)
+	}
+}

internal/auth/profile_metadata_readonly.go

@@ -0,0 +1,70 @@
+// Copyright 2026 Alibaba Group
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+package auth
+
+import (
+	"encoding/json"
+	"fmt"
+	"os"
+	"strings"
+)
+
+// ProfileMetadata is the minimal, non-sensitive identity projection exposed to
+// telemetry callers. It intentionally excludes profile names, client IDs,
+// organization names, token material, and credential status.
+type ProfileMetadata struct {
+	UserID   string
+	UserName string
+	CorpID   string
+}
+
+// ResolveProfileMetadataReadOnly resolves one identity exclusively from the
+// non-sensitive profiles.json metadata. It deliberately avoids auth locks,
+// token stores, Keychain access, migrations, quarantine renames, and writes.
+// A missing metadata file or an empty current profile returns (nil, nil).
+func ResolveProfileMetadataReadOnly(configDir, selector string) (*ProfileMetadata, error) {
+	data, err := profilesReadFile(ProfilesPath(configDir))
+	if err != nil {
+		if os.IsNotExist(err) {
+			return nil, nil
+		}
+		return nil, fmt.Errorf("read profile metadata: %w", err)
+	}
+
+	var cfg ProfilesConfig
+	if err := json.Unmarshal(data, &cfg); err != nil {
+		return nil, fmt.Errorf("parse profile metadata: %w", err)
+	}
+	if cfg.Version > profilesMaxVersion {
+		return nil, fmt.Errorf("profile metadata version %d is newer than supported version %d", cfg.Version, profilesMaxVersion)
+	}
+	normalizeProfilesConfig(&cfg)
+
+	selector = strings.TrimSpace(selector)
+	if selector == "" {
+		selector = strings.TrimSpace(cfg.CurrentProfile)
+		if selector == "" {
+			return nil, nil
+		}
+	}
+	profile, _, err := resolveProfileSelection("", &cfg, selector)
+	if err != nil {
+		return nil, err
+	}
+	return &ProfileMetadata{
+		UserID:   profile.UserID,
+		UserName: profile.UserName,
+		CorpID:   profile.CorpID,
+	}, nil
+}

internal/auth/profile_metadata_readonly_test.go

@@ -0,0 +1,97 @@
+// Copyright 2026 Alibaba Group
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+package auth
+
+import (
+	"errors"
+	"os"
+	"strings"
+	"testing"
+
+	"github.com/DingTalk-Real-AI/dingtalk-workspace-cli/internal/testseam"
+)
+
+func TestCrossPlatformCoverageResolveProfileMetadataReadOnly(t *testing.T) {
+	const metadata = `{
+  "version": 3,
+  "currentProfile": "corp-a:user-a",
+  "profiles": [
+    {"name":"alpha","corpId":"corp-a","userId":"user-a","userName":"Alice"},
+    {"name":"beta","corpId":"corp-b","userId":"user-b","userName":"Bob"}
+  ]
+}`
+	reads := 0
+	testseam.Swap(t, &profilesReadFile, func(path string) ([]byte, error) {
+		reads++
+		if !strings.HasSuffix(path, profilesJSONFile) {
+			t.Fatalf("metadata path = %q", path)
+		}
+		return []byte(metadata), nil
+	})
+
+	current, err := ResolveProfileMetadataReadOnly("/config", "")
+	if err != nil || current == nil || current.UserID != "user-a" || current.UserName != "Alice" || current.CorpID != "corp-a" {
+		t.Fatalf("current metadata profile = %#v, %v", current, err)
+	}
+	explicit, err := ResolveProfileMetadataReadOnly("/config", "beta")
+	if err != nil || explicit == nil || explicit.UserID != "user-b" || explicit.CorpID != "corp-b" {
+		t.Fatalf("explicit metadata profile = %#v, %v", explicit, err)
+	}
+	if reads != 2 {
+		t.Fatalf("profile metadata reads = %d, want 2", reads)
+	}
+}
+
+func TestCrossPlatformCoverageResolveProfileMetadataReadOnlyFailsClosed(t *testing.T) {
+	fail := errors.New("read failed")
+	for _, tc := range []struct {
+		name    string
+		read    func(string) ([]byte, error)
+		wantErr string
+	}{
+		{name: "missing", read: func(string) ([]byte, error) { return nil, os.ErrNotExist }},
+		{name: "read error", read: func(string) ([]byte, error) { return nil, fail }, wantErr: "read profile metadata"},
+		{name: "corrupt", read: func(string) ([]byte, error) { return []byte("{"), nil }, wantErr: "parse profile metadata"},
+		{name: "forward version", read: func(string) ([]byte, error) { return []byte(`{"version":999}`), nil }, wantErr: "newer than supported"},
+		{name: "no current", read: func(string) ([]byte, error) { return []byte(`{"version":3,"profiles":[]}`), nil }},
+	} {
+		t.Run(tc.name, func(t *testing.T) {
+			testseam.Swap(t, &profilesReadFile, tc.read)
+			testseam.Swap(t, &profilesRename, func(string, string) error {
+				t.Fatal("read-only metadata resolution attempted a quarantine rename")
+				return nil
+			})
+			got, err := ResolveProfileMetadataReadOnly("/config", "")
+			if tc.wantErr == "" {
+				if err != nil || got != nil {
+					t.Fatalf("read-only metadata = %#v, %v", got, err)
+				}
+				return
+			}
+			if err == nil || !strings.Contains(err.Error(), tc.wantErr) || got != nil {
+				t.Fatalf("read-only metadata = %#v, %v; want %q", got, err, tc.wantErr)
+			}
+		})
+	}
+}
+
+func TestCrossPlatformCoverageResolveProfileMetadataReadOnlyRejectsUnknownSelector(t *testing.T) {
+	testseam.Swap(t, &profilesReadFile, func(string) ([]byte, error) {
+		return []byte(`{"version":3,"currentProfile":"corp-a:user-a","profiles":[{"name":"alpha","corpId":"corp-a","userId":"user-a"}]}`), nil
+	})
+	profile, err := ResolveProfileMetadataReadOnly("/config", "missing")
+	if err == nil || profile != nil || !strings.Contains(err.Error(), "not found") {
+		t.Fatalf("unknown read-only profile = %#v, %v", profile, err)
+	}
+}

internal/cli/primary_param_schema_governance_test.go

@@ -0,0 +1,242 @@
+// Copyright 2026 Alibaba Group
+// Licensed under the Apache License, Version 2.0 (the "License");
+
+package cli
+
+import "testing"
+
+func TestPrimaryParamMigrationKeepsConceptDirectionCommandScoped(t *testing.T) {
+	concepts, err := LoadParamConcepts()
+	if err != nil {
+		t.Fatalf("LoadParamConcepts() error = %v", err)
+	}
+	if got := concepts.ByConcept["content_text"].CanonicalHint; got != "text" {
+		t.Fatalf("content_text canonical hint = %q, want unchanged global hint text", got)
+	}
+	if got := concepts.ByConcept["open_conversation_id"].CanonicalHint; got != "conversation-id" {
+		t.Fatalf("open_conversation_id canonical hint = %q, want unchanged global hint conversation-id", got)
+	}
+
+	content := concepts.ByConcept["content_text"]
+	docNode := concepts.ByConcept["doc_node_id"]
+	docAppend, problems := reduceLeafParamAliases(
+		"doc +doc-append",
+		realMap(
+			realFlag{name: "content"},
+			realFlag{name: "text", hidden: true},
+			realFlag{name: "doc"},
+		),
+		[]Concept{content, docNode},
+		CommandOverride{},
+	)
+	if len(problems) != 0 {
+		t.Fatalf("doc +doc-append reduction problems = %v", problems)
+	}
+	assertPrimaryParamAlias(t, docAppend, "body", "content")
+	assertPrimaryParamAlias(t, docAppend, "node", "doc")
+
+	insertOverride := primaryParamOverride(t, concepts, "doc block insert")
+	docInsert, problems := reduceLeafParamAliases(
+		"doc block insert",
+		realMap(
+			realFlag{name: "content"},
+			realFlag{name: "text", hidden: true},
+			realFlag{name: "parent-block"},
+			realFlag{name: "ref-block"},
+		),
+		[]Concept{content},
+		insertOverride,
+	)
+	if len(problems) != 0 {
+		t.Fatalf("doc block insert reduction problems = %v", problems)
+	}
+	assertPrimaryParamAlias(t, docInsert, "body", "content")
+
+	docUpdate, problems := reduceLeafParamAliases(
+		"doc block update",
+		realMap(realFlag{name: "content"}, realFlag{name: "text", hidden: true}),
+		[]Concept{content},
+		CommandOverride{},
+	)
+	if len(problems) != 0 {
+		t.Fatalf("doc block update reduction problems = %v", problems)
+	}
+	assertPrimaryParamAlias(t, docUpdate, "body", "content")
+
+	replyOverride := primaryParamOverride(t, concepts, "chat +messages-reply")
+	chatReply, problems := reduceLeafParamAliases(
+		"chat +messages-reply",
+		realMap(
+			realFlag{name: "group"},
+			realFlag{name: "conversation-id", hidden: true},
+			realFlag{name: "ref-msg-id"},
+		),
+		[]Concept{concepts.ByConcept["open_conversation_id"]},
+		replyOverride,
+	)
+	if len(problems) != 0 {
+		t.Fatalf("chat +messages-reply reduction problems = %v", problems)
+	}
+	for _, emitted := range []string{"chat", "chat-id", "open-conversation-id"} {
+		assertPrimaryParamAlias(t, chatReply, emitted, "group")
+	}
+	assertPrimaryParamAlias(t, chatReply, "msg-id", "ref-msg-id")
+	if chatReply.IsBlocked("group") {
+		t.Fatal("chat +messages-reply canonical --group must not remain blocked")
+	}
+}
+
+func TestPrimaryParamMigrationUpdatesReviewedFixturesAndMappingKeys(t *testing.T) {
+	concepts, err := LoadParamConcepts()
+	if err != nil {
+		t.Fatalf("LoadParamConcepts() error = %v", err)
+	}
+	fixtures := []ParamFixtureCase{
+		{Command: "doc +doc-append", Emitted: "text", Expect: "content", Via: "native:reviewed-compatibility-alias"},
+		{Command: "doc +doc-append", Emitted: "node", Expect: "doc", Via: "concept:doc_node_id"},
+		{Command: "doc block insert", Emitted: "text", Expect: "content", Via: "native:reviewed-compatibility-alias", Occ: 2},
+		{Command: "doc block update", Emitted: "text", Expect: "content", Via: "native:reviewed-compatibility-alias", Occ: 6},
+		{Command: "chat message send", Emitted: "file-path", Expect: "file", Via: "native:reviewed-compatibility-alias"},
+		{Command: "chat +messages-reply", Emitted: "conversation-id", Expect: "group", Via: "native:reviewed-compatibility-alias"},
+		{Command: "chat +messages-reply", Emitted: "chat", Expect: "group", Via: "concept:open_conversation_id"},
+		{Command: "chat +messages-reply", Emitted: "open-conversation-id", Expect: "group", Via: "concept:open_conversation_id"},
+	}
+	for _, want := range fixtures {
+		found := false
+		for _, got := range concepts.Fixture {
+			if got == want {
+				found = true
+				break
+			}
+		}
+		if !found {
+			t.Errorf("reviewed fixture missing %#v", want)
+		}
+	}
+
+	tests := []struct {
+		oldKey string
+		newKey string
+		reason string
+	}{
+		{"chat.reply_personal_message --text", "chat.reply_personal_message --content", "serialized into the aggregate content JSON string"},
+		{"chat.send_personal_message --file-path", "chat.send_personal_message --file", "local upload/preprocessing input"},
+		{"chat.send_personal_message --text", "chat.send_personal_message --content", "serialized into the aggregate content payload"},
+		{"doc.insert_document_block --text", "doc.insert_document_block --content", "aggregate convenience input used to build element"},
+		{"doc.update_document_block --text", "doc.update_document_block --content", "aggregate convenience input used to build element"},
+		{"todo.add_todo_attachment --file-path", "todo.add_todo_attachment --file", "local upload input used to construct attachmentList"},
+	}
+	for _, test := range tests {
+		if _, exists := reviewedSchemaParameterMappingExclusions[test.oldKey]; exists {
+			t.Errorf("stale mapping exclusion remains at %q", test.oldKey)
+		}
+		if got := reviewedSchemaParameterMappingExclusions[test.newKey]; got != test.reason {
+			t.Errorf("mapping exclusion %q reason = %q, want %q", test.newKey, got, test.reason)
+		}
+	}
+}
+
+func TestPrimaryParamMigrationPreservesDeliverySchemaSignatures(t *testing.T) {
+	tests := []struct {
+		path             string
+		primary          string
+		legacy           string
+		required         bool
+		cliRequired      bool
+		property         string
+		propertyFrom     string
+		format           string
+		requiredWhen     string
+		requiredWhenFrom string
+	}{
+		{path: "aisearch person", primary: "query", legacy: "keyword", required: true, property: "keyword", propertyFrom: "native_annotation", requiredWhenFrom: "default"},
+		{path: "chat message reply", primary: "content", legacy: "text", required: true, cliRequired: true, propertyFrom: "reviewed_mapping_exclusion", requiredWhenFrom: "default"},
+		{path: "chat message send", primary: "content", legacy: "text", propertyFrom: "reviewed_mapping_exclusion", requiredWhenFrom: "default"},
+		{path: "chat message send", primary: "file", legacy: "file-path", propertyFrom: "reviewed_mapping_exclusion", format: "file-path", requiredWhen: "msg-type is file or audio or video", requiredWhenFrom: "typed_parameter_metadata"},
+		{path: "chat message send-by-webhook", primary: "content", legacy: "text", required: true, cliRequired: true, property: "text", propertyFrom: "native_annotation", requiredWhenFrom: "default"},
+		{path: "chat +broadcast", primary: "content", legacy: "text", required: true, property: "text", propertyFrom: "native_annotation", requiredWhenFrom: "default"},
+		{path: "chat +dm", primary: "content", legacy: "text", required: true, property: "text", propertyFrom: "native_annotation", requiredWhenFrom: "default"},
+		{path: "chat +send-to-group", primary: "content", legacy: "text", required: true, property: "text", propertyFrom: "native_annotation", requiredWhenFrom: "default"},
+		{path: "chat +messages-send-by-bot", primary: "content", legacy: "text", required: true, property: "text", propertyFrom: "native_annotation", requiredWhenFrom: "default"},
+		{path: "chat +messages-batch-send-by-bot", primary: "content", legacy: "text", required: true, property: "text", propertyFrom: "native_annotation", requiredWhenFrom: "default"},
+		{path: "chat +messages-send-by-webhook", primary: "content", legacy: "text", required: true, property: "text", propertyFrom: "native_annotation", requiredWhenFrom: "default"},
+		{path: "chat +messages-reply", primary: "content", legacy: "text", required: true, property: "text", propertyFrom: "native_annotation", requiredWhenFrom: "default"},
+		{path: "chat +messages-reply", primary: "group", legacy: "conversation-id", required: true, property: "conversationId", propertyFrom: "native_annotation", requiredWhenFrom: "default"},
+		{path: "doc +doc-append", primary: "content", legacy: "text", required: true, property: "text", propertyFrom: "native_annotation", requiredWhenFrom: "default"},
+		{path: "doc block insert", primary: "content", legacy: "text", propertyFrom: "reviewed_mapping_exclusion", requiredWhenFrom: "default"},
+		{path: "doc block update", primary: "content", legacy: "text", propertyFrom: "reviewed_mapping_exclusion", requiredWhenFrom: "default"},
+		{path: "todo task add-attachment", primary: "file", legacy: "file-path", required: true, propertyFrom: "reviewed_mapping_exclusion", requiredWhenFrom: "default"},
+	}
+	for _, test := range tests {
+		t.Run(test.path+"/"+test.primary, func(t *testing.T) {
+			leaf, err := queryDeliverySchemaPayload([]string{test.path})
+			if err != nil {
+				t.Fatal(err)
+			}
+			parameters := schemaMap(leaf["parameters"])
+			parameter, exists := parameters[test.primary]
+			if !exists {
+				t.Fatalf("%s missing visible Primary --%s", test.path, test.primary)
+			}
+			if _, exists := parameters[test.legacy]; exists {
+				t.Fatalf("%s unexpectedly publishes hidden legacy --%s", test.path, test.legacy)
+			}
+			if got := parameter["type"]; got != "string" {
+				t.Errorf("%s --%s type = %#v, want string", test.path, test.primary, got)
+			}
+			if got := parameter["required"]; got != test.required {
+				t.Errorf("%s --%s required = %#v, want %v", test.path, test.primary, got, test.required)
+			}
+			if got, _ := parameter["cli_required"].(bool); got != test.cliRequired {
+				t.Errorf("%s --%s cli_required = %#v, want %v", test.path, test.primary, parameter["cli_required"], test.cliRequired)
+			}
+			if got, _ := parameter["property"].(string); got != test.property {
+				t.Errorf("%s --%s property = %q, want %q", test.path, test.primary, got, test.property)
+			}
+			if got := parameter["interface_type"]; got != nil && got != "" {
+				t.Errorf("%s --%s interface_type = %#v, want omitted/empty", test.path, test.primary, got)
+			}
+			if got, _ := parameter["format"].(string); got != test.format {
+				t.Errorf("%s --%s format = %q, want %q", test.path, test.primary, got, test.format)
+			}
+			if got, _ := parameter["required_when"].(string); got != test.requiredWhen {
+				t.Errorf("%s --%s required_when = %q, want %q", test.path, test.primary, got, test.requiredWhen)
+			}
+
+			provenance := schemaMap(parameter["field_provenance"])
+			propertyProvenance := provenance["property"]
+			if got := propertyProvenance["source"]; got != test.propertyFrom {
+				t.Errorf("%s --%s property source = %#v, want %s", test.path, test.primary, got, test.propertyFrom)
+			}
+			typeProvenance := provenance["type"]
+			if got := typeProvenance["source"]; got != "cobra_flag_type" {
+				t.Errorf("%s --%s type source = %#v, want cobra_flag_type", test.path, test.primary, got)
+			}
+			requiredWhenProvenance := provenance["required_when"]
+			if got := requiredWhenProvenance["source"]; got != test.requiredWhenFrom {
+				t.Errorf("%s --%s required_when source = %#v, want %s", test.path, test.primary, got, test.requiredWhenFrom)
+			}
+		})
+	}
+}
+
+func primaryParamOverride(t *testing.T, concepts ParamConcepts, path string) CommandOverride {
+	t.Helper()
+	for _, override := range concepts.Overrides {
+		if override.CommandPath == path {
+			return override
+		}
+	}
+	t.Fatalf("command override %q not found", path)
+	return CommandOverride{}
+}
+
+func assertPrimaryParamAlias(t *testing.T, entry *ParamAliasEntry, emitted, want string) {
+	t.Helper()
+	if entry == nil {
+		t.Fatalf("alias entry is nil; want %s -> %s", emitted, want)
+	}
+	if got, ok := entry.ResolveAlias(emitted); !ok || got != want {
+		t.Fatalf("ResolveAlias(%q) = %q (ok=%v), want %q", emitted, got, ok, want)
+	}
+}

internal/cli/runtime_schema.go

@@ -1058,7 +1058,7 @@ var schemaCompactPayloadKeys = map[string]bool{
 	"agent_summary": true, "description": true,
 	"effect": true, "risk": true, "confirmation": true, "idempotency": true,
 	"interface_mode": true, "availability": true, "interface_reason": true,
-	"parameters": true, "constraints": true, "positionals": true, "dry_run": true,
+	"parameters": true, "constraints": true, "positionals": true, "dry_run": true, "wait": true,
 	"result": true, "pagination": true,
 	"examples": true, "use_when": true, "avoid_when": true,
 }

internal/cli/schema_catalog_structure.go

@@ -81,6 +81,7 @@ var schemaCatalogToolOptionalKeys = []string{
 	"pagination",
 	"positionals",
 	"result",
+	"wait",
 }
 
 var schemaCatalogToolEnums = map[string][]string{

internal/cli/schema_constraints_runtime.go

@@ -68,9 +68,9 @@ func init() {
 	registerRequireOneOf("devdoc.search_open_platform_docs_rag", "query", "keyword")
 	registerRequireOneOf("event.consume", "event_key", "subscribe-id")
 	registerExclusiveOneOf("event.stop", "all", "subscribe_id")
-	registerRequireOneOf("doc.insert_document_block", "text", "heading", "element")
+	registerRequireOneOf("doc.insert_document_block", "content", "heading", "element")
 	registerExclusiveOneOf("doc.update_document", "content", "content-file")
-	registerRequireOneOf("doc.update_document_block", "text", "heading", "element")
+	registerRequireOneOf("doc.update_document_block", "content", "heading", "element")
 	registerRequireOneOf("pat.batch_grant", "scope", "product", "products", "domain", "domains", "recommend")
 	registerRequireOneOf("mail.search_mail_users", "keyword", "employee-no")
 	// --body is a hidden compatibility alias for the public --content flag.

internal/cli/schema_contract_model.go

@@ -60,6 +60,7 @@ type ToolSpec struct {
 	Constraints     RuntimeSchemaConstraints
 	Positionals     []contract.RuntimeSchemaPositional
 	DryRun          *contract.DryRunSpec
+	Wait            *contract.WaitSpec
 	Result          *contract.ResultSpec
 	Pagination      *contract.PaginationSpec
 	Safety          contract.SafetySpec
@@ -135,6 +136,7 @@ type RuntimeToolSpecInput struct {
 	Constraints     RuntimeSchemaConstraints
 	Positionals     []contract.RuntimeSchemaPositional
 	DryRun          *contract.DryRunSpec
+	Wait            *contract.WaitSpec
 	Result          *contract.ResultSpec
 	Pagination      *contract.PaginationSpec
 	Safety          contract.SafetySpec
@@ -542,6 +544,11 @@ func (t ToolSpec) Validate() error {
 			return err
 		}
 	}
+	if t.Wait != nil {
+		if err := t.Wait.Validate(id.CanonicalPath); err != nil {
+			return err
+		}
+	}
 	if t.Result != nil {
 		if _, err := contract.NormalizeResultSpec(t.Result, id.CanonicalPath); err != nil {
 			return err
@@ -744,6 +751,16 @@ func (t ToolSpec) normalized() ToolSpec {
 		dryRun.PreviewKind = strings.TrimSpace(dryRun.PreviewKind)
 		out.DryRun = &dryRun
 	}
+	if t.Wait != nil {
+		// NormalizeWaitSpec is the single canonical form shared with the
+		// declaration path: trimmed status values, duplicate/conflict
+		// rejection, defensive copy. Invalid declarations are rejected by
+		// ToolSpec.Validate below, which runs the same normalization
+		// through WaitSpec.Validate.
+		if wait, err := contract.NormalizeWaitSpec(t.Wait, id.CanonicalPath); err == nil {
+			out.Wait = wait
+		}
+	}
 	if t.Result != nil {
 		result, err := contract.NormalizeResultSpec(t.Result, id.CanonicalPath)
 		if err == nil {
@@ -982,6 +999,10 @@ func (t ToolSpec) ToPayload() (map[string]any, error) {
 		value, _ := typedJSONValue(t.DryRun)
 		payload["dry_run"] = value
 	}
+	if t.Wait != nil {
+		value, _ := typedJSONValue(t.Wait)
+		payload["wait"] = value
+	}
 	if t.Result != nil {
 		value, _ := typedJSONValue(t.Result)
 		payload["result"] = value
@@ -1127,21 +1148,37 @@ func putRawJSON(payload map[string]any, key string, raw json.RawMessage) error {
 	return nil
 }
 
+// rawJSONValue decodes a JSON value that may come from an untrusted source, so
+// it validates before decoding. Callers holding output that json.Marshal just
+// produced should use typedJSONValue instead of paying the validation scan.
 func rawJSONValue(raw json.RawMessage) (any, error) {
 	if !json.Valid(raw) {
 		return nil, fmt.Errorf("invalid JSON value")
 	}
+	return decodeValidJSONValue(raw), nil
+}
+
+// decodeValidJSONValue decodes JSON whose validity the caller has already
+// established, either by json.Valid or by having just marshaled it. Decode
+// errors are unreachable under that precondition and are therefore discarded,
+// exactly as this path behaved when the decode was inlined into rawJSONValue.
+func decodeValidJSONValue(raw json.RawMessage) any {
 	decoder := json.NewDecoder(bytes.NewReader(raw))
 	decoder.UseNumber()
 	var value any
 	_ = decoder.Decode(&value)
-	return value, nil
+	return value
 }
 
+// typedJSONValue projects a typed value into the generic JSON shape the payload
+// renderers consume. json.Marshal output is valid by construction, so this path
+// decodes it directly: routing through rawJSONValue re-scanned every marshaled
+// document with json.Valid, which measured ~34% of Schema Catalog assembly time
+// across the 1121-tool set (26.0s -> 17.2s for the internal/app schema suite).
 func typedJSONValue(value any) (any, error) {
 	data, err := json.Marshal(value)
 	if err != nil {
 		return nil, err
 	}
-	return rawJSONValue(data)
+	return decodeValidJSONValue(data), nil
 }

internal/cli/schema_contract_model_test.go

@@ -974,3 +974,57 @@ func TestFinalProvenanceCoverageDoesNotInventOptionalInterfaceReason(t *testing.
 		t.Fatalf("optional local interface_reason should not require invented provenance: %v", err)
 	}
 }
+
+func TestToolSpecWaitCapabilityIsPositiveOnly(t *testing.T) {
+	base := RuntimeToolSpecInput{Identity: contract.ToolIdentitySpec{
+		ProductID: "sample",
+		Name:      "waitrun",
+		CLIName:   "waitrun",
+		CLIPath:   "sample waitrun",
+	}}
+
+	withoutCapability, err := ToolSpecFromRuntime(base)
+	if err != nil {
+		t.Fatalf("ToolSpecFromRuntime() error = %v", err)
+	}
+	payload, err := withoutCapability.ToPayload()
+	if err != nil {
+		t.Fatalf("ToPayload() error = %v", err)
+	}
+	if _, ok := payload["wait"]; ok {
+		t.Fatalf("nil capability unexpectedly emitted wait: %#v", payload["wait"])
+	}
+
+	base.Wait = &contract.WaitSpec{Mode: "webhook"}
+	if _, err := ToolSpecFromRuntime(base); err == nil || !strings.Contains(err.Error(), "unknown mode") {
+		t.Fatalf("invalid mode error = %v", err)
+	}
+	base.Wait = &contract.WaitSpec{Mode: contract.WaitModeEvent, StatusQuery: "status", Terminal: map[string]contract.ResultOutcome{"DONE": contract.ResultOutcomeSuccess}}
+	if _, err := ToolSpecFromRuntime(base); err == nil || !strings.Contains(err.Error(), "requires event_key") {
+		t.Fatalf("event mode body error = %v", err)
+	}
+
+	base.Wait = &contract.WaitSpec{
+		Mode:               contract.WaitModePoll,
+		PollCommand:        "sample status get",
+		StatusQuery:        "result.status",
+		Terminal:           map[string]contract.ResultOutcome{"COMPLETED": contract.ResultOutcomeSuccess},
+		PendingValues:      []string{"NEW"},
+		DefaultTimeoutSecs: 120,
+	}
+	withCapability, err := ToolSpecFromRuntime(base)
+	if err != nil {
+		t.Fatalf("ToolSpecFromRuntime(valid wait) error = %v", err)
+	}
+	if withCapability.Wait == nil || withCapability.Wait.Mode != contract.WaitModePoll {
+		t.Fatalf("wait capability lost through normalization: %#v", withCapability.Wait)
+	}
+	payload, err = withCapability.ToPayload()
+	if err != nil {
+		t.Fatalf("ToPayload(valid wait) error = %v", err)
+	}
+	wait := payload["wait"].(map[string]any)
+	if wait["mode"] != contract.WaitModePoll || wait["poll_command"] != "sample status get" {
+		t.Fatalf("wait payload=%#v", wait)
+	}
+}

internal/cli/schema_parameter_mapping_ledger.go

@@ -269,20 +269,20 @@ var reviewedSchemaParameterMappingExclusions = map[string]string{
 	"chat.update_text_emotion --conversation-id":                         "Reviewed unpinned adapter: chat.update_text_emotion has no singular pinned interface_ref; --conversation-id is a CLI wrapper input and does not publish a direct interface property.",
 	"chat.reply_personal_message --ref-msg-id":                           "serialized into the aggregate content JSON string",
 	"chat.reply_personal_message --ref-sender":                           "resolved then serialized into the aggregate content JSON string",
-	"chat.reply_personal_message --text":                                 "serialized into the aggregate content JSON string",
+	"chat.reply_personal_message --content":                              "serialized into the aggregate content JSON string",
 	"chat.search_groups --cursor":                                        "Reviewed unpinned adapter: chat.search_groups has no singular pinned interface_ref; --cursor is a CLI wrapper input and does not publish a direct interface property.",
 	"chat.search_groups --exclude-muted":                                 "Reviewed unpinned adapter: chat.search_groups has no singular pinned interface_ref; --exclude-muted is a CLI wrapper input and does not publish a direct interface property.",
 	"chat.search_groups --limit":                                         "Reviewed unpinned adapter: chat.search_groups has no singular pinned interface_ref; --limit is a CLI wrapper input and does not publish a direct interface property.",
 	"chat.search_messages --user":                                        "conditional wrapper: parseCSVValues + appendChatIDArgs routes each supplied identifier to senderUserIds or senderOpenDingTakIds according to its runtime ID shape; there is no single RPC property for this flag",
 	"chat.search_messages --users":                                       "conditional wrapper/alias of --user: parseCSVValues + appendChatIDArgs routes each supplied identifier to senderUserIds or senderOpenDingTakIds according to its runtime ID shape; there is no single RPC property for this flag",
 	"chat.send_personal_message --contact-id":                            "serialized into the aggregate content payload",
-	"chat.send_personal_message --file-path":                             "local upload/preprocessing input",
+	"chat.send_personal_message --file":                                  "local upload/preprocessing input",
 	"chat.send_personal_message --latitude":                              "serialized into the aggregate content payload",
 	"chat.send_personal_message --location-name":                         "serialized into the aggregate content payload",
 	"chat.send_personal_message --longitude":                             "serialized into the aggregate content payload",
 	"chat.send_personal_message --map-thumbnail-url":                     "serialized into the aggregate content payload",
 	"chat.send_personal_message --media-id":                              "serialized into the aggregate content payload",
-	"chat.send_personal_message --text":                                  "serialized into the aggregate content payload",
+	"chat.send_personal_message --content":                               "serialized into the aggregate content payload",
 	"chat.send_personal_message --title":                                 "serialized into the aggregate content payload",
 	"chat.send_personal_message --user":                                  "resolves a userId to receiverOpenDingTalkId when possible and otherwise uses the unpinned receiverUserId compatibility property",
 	"chat.set_group_member_mute_list --user":                             "conditional wrapper/alias of --users: userIds are remotely resolved to openDingTalkIds outside dry-run and otherwise may be sent as uids; openDingTalkId inputs are sent as openDingTalkIds, so one exact property would misdescribe execution",
@@ -406,7 +406,7 @@ var reviewedSchemaParameterMappingExclusions = map[string]string{
 	"doc.insert_document_block --fix-jsonml":                             "local JSONML normalization control",
 	"doc.insert_document_block --heading":                                "aggregate convenience input used to build element",
 	"doc.insert_document_block --level":                                  "aggregate convenience input used to build element",
-	"doc.insert_document_block --text":                                   "aggregate convenience input used to build element",
+	"doc.insert_document_block --content":                                "aggregate convenience input used to build element",
 	"doc.list_document_blocks --block-id":                                "runtime extension sends blockId, which is absent from the pinned list_document_blocks metadata",
 	"doc.reply_comment --mentioned-open-conversation-id":                 "Runtime extension sends mentionedOpenConversationIds, which is absent from the immutable pinned reply_comment metadata at its declared source revision.",
 	"doc.style_background_clear --node":                                  "Reviewed unpinned adapter: doc.style_background_clear has no singular pinned interface_ref; --node is a CLI wrapper input and does not publish a direct interface property.",
@@ -438,7 +438,7 @@ var reviewedSchemaParameterMappingExclusions = map[string]string{
 	"doc.update_document_block --fix-jsonml":                             "local JSONML normalization control",
 	"doc.update_document_block --heading":                                "aggregate convenience input used to build element",
 	"doc.update_document_block --level":                                  "aggregate convenience input used to build element",
-	"doc.update_document_block --text":                                   "aggregate convenience input used to build element",
+	"doc.update_document_block --content":                                "aggregate convenience input used to build element",
 	"doc.version_list --cursor":                                          "Reviewed unpinned adapter: doc.version_list has no singular pinned interface_ref; --cursor is a CLI wrapper input and does not publish a direct interface property.",
 	"doc.version_list --limit":                                           "Reviewed unpinned adapter: doc.version_list has no singular pinned interface_ref; --limit is a CLI wrapper input and does not publish a direct interface property.",
 	"doc.version_list --node":                                            "Reviewed unpinned adapter: doc.version_list has no singular pinned interface_ref; --node is a CLI wrapper input and does not publish a direct interface property.",
@@ -659,7 +659,7 @@ var reviewedSchemaParameterMappingExclusions = map[string]string{
 	"sheet.write_image --file":                                           "local upload input used to obtain resourceId/resourceUrl",
 	"sheet.write_image --mime-type":                                      "local upload metadata",
 	"sheet.write_image --name":                                           "local upload metadata",
-	"todo.add_todo_attachment --file-path":                               "local upload input used to construct attachmentList",
+	"todo.add_todo_attachment --file":                                    "local upload input used to construct attachmentList",
 	"todo.get_user_todos_in_current_org --query-all":                     "Local route selector: the default path calls get_user_todos_in_current_org, while --query-all switches to get_user_todos; it is not a property of the pinned default RPC.",
 	"todo.list_todo_attachment --task-id":                                "Reviewed unpinned adapter: --task-id is nested under todoAttachmentListRequest at runtime, while the immutable pinned MCP snapshot has no interface_ref for todo.list_todo_attachment.",
 	"wiki.create_wikiSpace --icon":                                       "runtime extension sends icon, which is absent from the pinned create_wikiSpace metadata",

internal/cli/schema_parameters_runtime.go

@@ -77,7 +77,7 @@ func cloneRuntimeSchemaStringMap(source map[string]string) map[string]string {
 
 func init() {
 	RegisterRuntimeSchemaParameterMetadata("aisearch.enterprise_person_search", RuntimeSchemaParameterMetadata{
-		Required: []string{"keyword"},
+		Required: []string{"query"},
 	})
 	RegisterRuntimeSchemaParameterMetadata("aitable.export_data", RuntimeSchemaParameterMetadata{
 		RequiredWhen: map[string]string{
@@ -146,8 +146,8 @@ func init() {
 	})
 	RegisterRuntimeSchemaParameterMetadata("chat.send_personal_message", RuntimeSchemaParameterMetadata{
 		RequiredWhen: map[string]string{
-			"media-id":  "msg-type is image",
-			"file-path": "msg-type is file or audio or video",
+			"media-id": "msg-type is image",
+			"file":     "msg-type is file or audio or video",
 		},
 		Examples: map[string]string{"media-id": "@lADP_schema_smoke"},
 	})

internal/cli/schema_runtime_registry.go

@@ -354,6 +354,7 @@ func runtimeToolSpecFromContractFinal(entry runtimeSchemaEntry, final contract.C
 		Constraints:     constraints,
 		Positionals:     positionals,
 		DryRun:          final.DryRun,
+		Wait:            final.Wait,
 		Result:          result,
 		Pagination:      pagination,
 		Safety:          safety,

internal/cli/schema_snapshot_adapter.go

@@ -65,6 +65,7 @@ type schemaToolWire struct {
 	Constraints         RuntimeSchemaConstraints            `json:"constraints"`
 	Positionals         []contract.RuntimeSchemaPositional  `json:"positionals"`
 	DryRun              *contract.DryRunSpec                `json:"dry_run"`
+	Wait                *contract.WaitSpec                  `json:"wait"`
 	Result              *contract.ResultSpec                `json:"result"`
 	Pagination          *contract.PaginationSpec            `json:"pagination"`
 	Effect              string                              `json:"effect"`
@@ -269,6 +270,7 @@ func schemaToolSpecFromWire(wire schemaToolWire) (ToolSpec, error) {
 		Constraints:    wire.Constraints,
 		Positionals:    wire.Positionals,
 		DryRun:         wire.DryRun,
+		Wait:           wire.Wait,
 		Result:         wire.Result,
 		Pagination:     wire.Pagination,
 		Safety: contract.SafetySpec{

internal/corecmd/contract/final.go

@@ -30,6 +30,7 @@ type ContractFinalPayload struct {
 	Parameters  []ParamDecl
 	Safety      *SafetySpec
 	DryRun      *DryRunSpec
+	Wait        *WaitSpec
 	Result      *ResultSpec
 	Pagination  *PaginationSpec
 	Interface   *InterfaceSpec

internal/corecmd/contract/types.go

@@ -62,6 +62,155 @@ type DryRunSpec struct {
 	RemoteReads bool   `json:"remote_reads,omitempty"`
 }
 
+// Wait modes. Poll executes the leaf's WaitPoll hook on a cadence. Event
+// consumes the leaf's WaitEvents push stream and correlates events to the
+// accepted resource. Auto prefers the event stream and falls back to polling
+// when the stream ends before a terminal status.
+const (
+	WaitModePoll  = "poll"
+	WaitModeEvent = "event"
+	WaitModeAuto  = "auto"
+)
+
+// WaitSpec is a positive capability declaration for terminal-state waiting
+// (approval flows, async exports, batch jobs). A nil ToolSpec.Wait means the
+// command has not declared reviewed --wait support; the flag is not
+// registered and the Schema does not publish the capability.
+//
+// Like DryRunSpec, the object is one atomic contract field: Schema only
+// projects the reviewed capability; runtime execution stays owned by the
+// command runner through the leaf's WaitPoll / WaitEvents hooks. PollCommand
+// names the read command that observes status — it is a declared,
+// catalog-visible fact (the same command an agent would poll manually), not
+// a framework-owned invocation: how one poll or event subscription executes
+// is decided by the leaf.
+type WaitSpec struct {
+	Mode          string                   `json:"mode"`
+	PollCommand   string                   `json:"poll_command,omitempty"`
+	StatusQuery   string                   `json:"status_query"`
+	Terminal      map[string]ResultOutcome `json:"terminal"`
+	PendingValues []string                 `json:"pending_values,omitempty"`
+	// EventKey is the push channel key the WaitEvents hook subscribes to
+	// (event/auto modes). Declared for the catalog; the transport stays
+	// leaf-owned.
+	EventKey string `json:"event_key,omitempty"`
+	// MatchField is the event-document path holding the resource identifier
+	// (event/auto modes); its value must equal the ResourceQuery resolution
+	// of the accepted result.
+	MatchField string `json:"match_field,omitempty"`
+	// ResourceQuery is the dotted path into the accepted result data that
+	// yields the resource identifier correlated against MatchField
+	// (event/auto modes).
+	ResourceQuery string `json:"resource_query,omitempty"`
+	// DefaultTimeoutSecs is the reviewed default for --wait-timeout. Zero
+	// means the framework default (300s); the user flag always wins.
+	DefaultTimeoutSecs int `json:"default_timeout_secs"`
+}
+
+// Validate checks mode requirements and the terminal/pending status maps.
+// Unknown terminal outcomes, unknown modes, and mode/body mismatches fail at
+// declaration so a malformed wait capability cannot reach the wire.
+// Validation delegates to NormalizeWaitSpec so the acceptance rules can never
+// drift from the normalization the wire and the runtime wait engine share.
+func (w WaitSpec) Validate(canonical string) error {
+	_, err := NormalizeWaitSpec(&w, canonical)
+	return err
+}
+
+// NormalizeWaitSpec returns a validated, canonical, defensively copied wait
+// contract. It is shared by declaration (corecmd.New / AttachContract),
+// ToolSpec, and snapshot paths, mirroring NormalizeResultSpec. Status values
+// are trimmed into their wire form: the wait engine compares backend
+// statuses verbatim against these tables, so a padded declaration
+// (" processing ") would publish a Schema that its own runtime treats as an
+// unknown status. Values collapsing onto one value after trimming (duplicate
+// pending values, duplicate terminal keys, terminal/pending conflicts) are
+// rejected instead of silently merged.
+func NormalizeWaitSpec(in *WaitSpec, canonical string) (*WaitSpec, error) {
+	if in == nil {
+		return nil, nil
+	}
+	canonical = defaultString(strings.TrimSpace(canonical), "<unknown>")
+	out := &WaitSpec{
+		Mode:               strings.TrimSpace(in.Mode),
+		PollCommand:        strings.TrimSpace(in.PollCommand),
+		StatusQuery:        strings.TrimSpace(in.StatusQuery),
+		EventKey:           strings.TrimSpace(in.EventKey),
+		MatchField:         strings.TrimSpace(in.MatchField),
+		ResourceQuery:      strings.TrimSpace(in.ResourceQuery),
+		DefaultTimeoutSecs: in.DefaultTimeoutSecs,
+	}
+	if out.Mode == "" {
+		return nil, fmt.Errorf("schema tool %s wait has no mode", canonical)
+	}
+	switch out.Mode {
+	case WaitModePoll, WaitModeEvent, WaitModeAuto:
+	default:
+		return nil, fmt.Errorf("schema tool %s wait has unknown mode %q", canonical, out.Mode)
+	}
+	needsPoll := out.Mode == WaitModePoll || out.Mode == WaitModeAuto
+	if needsPoll && out.PollCommand == "" {
+		return nil, fmt.Errorf("schema tool %s wait mode %s requires poll_command", canonical, out.Mode)
+	}
+	needsEvent := out.Mode == WaitModeEvent || out.Mode == WaitModeAuto
+	if needsEvent {
+		if out.EventKey == "" {
+			return nil, fmt.Errorf("schema tool %s wait mode %s requires event_key", canonical, out.Mode)
+		}
+		if out.MatchField == "" {
+			return nil, fmt.Errorf("schema tool %s wait mode %s requires match_field", canonical, out.Mode)
+		}
+		if out.ResourceQuery == "" {
+			return nil, fmt.Errorf("schema tool %s wait mode %s requires resource_query", canonical, out.Mode)
+		}
+	}
+	if out.StatusQuery == "" {
+		return nil, fmt.Errorf("schema tool %s wait mode %s requires status_query", canonical, out.Mode)
+	}
+	if len(in.Terminal) == 0 {
+		return nil, fmt.Errorf("schema tool %s wait has no terminal states", canonical)
+	}
+	out.Terminal = make(map[string]ResultOutcome, len(in.Terminal))
+	for status, outcome := range in.Terminal {
+		status = strings.TrimSpace(status)
+		if status == "" {
+			return nil, fmt.Errorf("schema tool %s wait has a blank terminal status", canonical)
+		}
+		if _, dup := out.Terminal[status]; dup {
+			return nil, fmt.Errorf("schema tool %s wait has duplicate terminal status %q", canonical, status)
+		}
+		// Terminal states must close into success or failure. Pending and
+		// partial are not wait outcomes: pending is expressed through
+		// timeout, and partial requires the typed multi-status payload only
+		// the leaf can construct.
+		if outcome != ResultOutcomeSuccess && outcome != ResultOutcomeFailure {
+			return nil, fmt.Errorf(
+				"schema tool %s wait terminal status %q must map to success or failure, got %q",
+				canonical, status, outcome)
+		}
+		out.Terminal[status] = outcome
+	}
+	seenPending := make(map[string]bool, len(in.PendingValues))
+	for _, value := range in.PendingValues {
+		value = strings.TrimSpace(value)
+		if value == "" {
+			return nil, fmt.Errorf("schema tool %s wait has a blank pending value", canonical)
+		}
+		if _, conflict := out.Terminal[value]; conflict {
+			return nil, fmt.Errorf("schema tool %s wait status %q is both terminal and pending", canonical, value)
+		}
+		if seenPending[value] {
+			return nil, fmt.Errorf("schema tool %s wait has duplicate pending value %q", canonical, value)
+		}
+		seenPending[value] = true
+		out.PendingValues = append(out.PendingValues, value)
+	}
+	if in.DefaultTimeoutSecs < 0 {
+		return nil, fmt.Errorf("schema tool %s wait default_timeout_secs must be >= 0", canonical)
+	}
+	return out, nil
+}
+
 // ResultOutcome is one closed unified-output envelope outcome.
 type ResultOutcome string
 

internal/corecmd/contract/wait_spec_test.go

@@ -0,0 +1,227 @@
+// Copyright 2026 Alibaba Group
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+package contract
+
+import "testing"
+
+func TestWaitSpecValidateAcceptsReviewedShapes(t *testing.T) {
+	cases := []WaitSpec{
+		{
+			Mode:        WaitModePoll,
+			PollCommand: "oa approval-instance get",
+			StatusQuery: "result.status",
+			Terminal: map[string]ResultOutcome{
+				"COMPLETED": ResultOutcomeSuccess,
+				"REJECTED":  ResultOutcomeFailure,
+			},
+			PendingValues:      []string{"NEW", "RUNNING"},
+			DefaultTimeoutSecs: 600,
+		},
+	}
+	for i, spec := range cases {
+		if err := spec.Validate("sample.tool"); err != nil {
+			t.Fatalf("case %d: unexpected error: %v", i, err)
+		}
+	}
+}
+
+func TestWaitSpecValidateRejectsMalformedShapes(t *testing.T) {
+	terminal := map[string]ResultOutcome{"COMPLETED": ResultOutcomeSuccess}
+	cases := map[string]WaitSpec{
+		"no mode": {
+			Terminal: terminal,
+		},
+		"unknown mode": {
+			Mode:     "webhook",
+			Terminal: terminal,
+		},
+		"poll without poll_command": {
+			Mode:        WaitModePoll,
+			StatusQuery: "status",
+			Terminal:    terminal,
+		},
+		"poll without status_query": {
+			Mode:        WaitModePoll,
+			PollCommand: "oa approval-instance get",
+			Terminal:    terminal,
+		},
+		"event without event_key": {
+			Mode:          WaitModeEvent,
+			MatchField:    "process_instance_id",
+			ResourceQuery: "id",
+			StatusQuery:   "result.status",
+			Terminal:      terminal,
+		},
+		"event without match_field": {
+			Mode:          WaitModeEvent,
+			EventKey:      "bpms_instance_change",
+			ResourceQuery: "id",
+			StatusQuery:   "result.status",
+			Terminal:      terminal,
+		},
+		"event without resource_query": {
+			Mode:        WaitModeEvent,
+			EventKey:    "bpms_instance_change",
+			MatchField:  "process_instance_id",
+			StatusQuery: "result.status",
+			Terminal:    terminal,
+		},
+		"auto missing poll_command": {
+			Mode:          WaitModeAuto,
+			EventKey:      "export_finished",
+			MatchField:    "job_id",
+			ResourceQuery: "job_id",
+			StatusQuery:   "status",
+			Terminal:      terminal,
+		},
+		"no terminal states": {
+			Mode:        WaitModePoll,
+			PollCommand: "oa approval-instance get",
+			StatusQuery: "status",
+		},
+		"blank terminal status": {
+			Mode:        WaitModePoll,
+			PollCommand: "oa approval-instance get",
+			StatusQuery: "status",
+			Terminal:    map[string]ResultOutcome{" ": ResultOutcomeSuccess},
+		},
+		"terminal outcome pending": {
+			Mode:        WaitModePoll,
+			PollCommand: "oa approval-instance get",
+			StatusQuery: "status",
+			Terminal:    map[string]ResultOutcome{"COMPLETED": ResultOutcomePending, "REJECTED": ResultOutcomeFailure},
+		},
+		"terminal outcome partial": {
+			Mode:        WaitModePoll,
+			PollCommand: "oa approval-instance get",
+			StatusQuery: "status",
+			Terminal:    map[string]ResultOutcome{"COMPLETED": ResultOutcomePartialFailure, "REJECTED": ResultOutcomeFailure},
+		},
+		"terminal outcome outside closed set": {
+			Mode:        WaitModePoll,
+			PollCommand: "oa approval-instance get",
+			StatusQuery: "status",
+			Terminal:    map[string]ResultOutcome{"COMPLETED": ResultOutcome("explosion")},
+		},
+		"only pending terminal outcome": {
+			Mode:        WaitModePoll,
+			PollCommand: "oa approval-instance get",
+			StatusQuery: "status",
+			Terminal:    map[string]ResultOutcome{"NEW": ResultOutcomePending},
+		},
+		"status both terminal and pending": {
+			Mode:          WaitModePoll,
+			PollCommand:   "oa approval-instance get",
+			StatusQuery:   "status",
+			Terminal:      terminal,
+			PendingValues: []string{"COMPLETED"},
+		},
+		"blank pending value": {
+			Mode:          WaitModePoll,
+			PollCommand:   "oa approval-instance get",
+			StatusQuery:   "status",
+			Terminal:      terminal,
+			PendingValues: []string{"  "},
+		},
+		"negative timeout default": {
+			Mode:               WaitModePoll,
+			PollCommand:        "oa approval-instance get",
+			StatusQuery:        "status",
+			Terminal:           terminal,
+			DefaultTimeoutSecs: -1,
+		},
+	}
+	for name, spec := range cases {
+		if err := spec.Validate("sample.tool"); err == nil {
+			t.Fatalf("%s: expected error, got nil", name)
+		}
+	}
+}
+
+func TestNormalizeWaitSpecTrimsStatusValuesIntoWireForm(t *testing.T) {
+	in := &WaitSpec{
+		Mode:               "  poll  ",
+		PollCommand:        " oa approval-instance get ",
+		StatusQuery:        " result.status ",
+		Terminal:           map[string]ResultOutcome{" COMPLETED ": ResultOutcomeSuccess, "REJECTED": ResultOutcomeFailure},
+		PendingValues:      []string{" NEW ", "RUNNING"},
+		DefaultTimeoutSecs: 60,
+	}
+	out, err := NormalizeWaitSpec(in, "sample.tool")
+	if err != nil {
+		t.Fatalf("NormalizeWaitSpec() error = %v", err)
+	}
+	if out.Mode != WaitModePoll || out.PollCommand != "oa approval-instance get" || out.StatusQuery != "result.status" {
+		t.Fatalf("normalized scalars: %#v", out)
+	}
+	if len(out.Terminal) != 2 {
+		t.Fatalf("terminal=%#v, want two trimmed keys", out.Terminal)
+	}
+	if got := out.Terminal["COMPLETED"]; got != ResultOutcomeSuccess {
+		t.Fatalf("terminal[COMPLETED]=%q, want success (key must be trimmed)", got)
+	}
+	if _, padded := out.Terminal[" COMPLETED "]; padded {
+		t.Fatal("padded terminal key survived normalization")
+	}
+	for i, want := range []string{"NEW", "RUNNING"} {
+		if out.PendingValues[i] != want {
+			t.Fatalf("pending[%d]=%q, want %q", i, out.PendingValues[i], want)
+		}
+	}
+	// The input declaration must stay untouched (defensive copy).
+	if _, padded := in.Terminal[" COMPLETED "]; !padded {
+		t.Fatal("NormalizeWaitSpec mutated its input terminal map")
+	}
+	if in.PendingValues[0] != " NEW " {
+		t.Fatal("NormalizeWaitSpec mutated its input pending values")
+	}
+}
+
+func TestNormalizeWaitSpecRejectsDuplicatesAndConflictsAfterTrim(t *testing.T) {
+	terminal := map[string]ResultOutcome{"COMPLETED": ResultOutcomeSuccess}
+	cases := map[string]*WaitSpec{
+		"terminal keys collapsing after trim": {
+			Mode:        WaitModePoll,
+			PollCommand: "oa approval-instance get",
+			StatusQuery: "status",
+			Terminal:    map[string]ResultOutcome{"COMPLETED": ResultOutcomeSuccess, " COMPLETED ": ResultOutcomeFailure},
+		},
+		"pending values collapsing after trim": {
+			Mode:          WaitModePoll,
+			PollCommand:   "oa approval-instance get",
+			StatusQuery:   "status",
+			Terminal:      terminal,
+			PendingValues: []string{"NEW", " NEW "},
+		},
+		"terminal/pending conflict hidden by padding": {
+			Mode:          WaitModePoll,
+			PollCommand:   "oa approval-instance get",
+			StatusQuery:   "status",
+			Terminal:      terminal,
+			PendingValues: []string{" COMPLETED "},
+		},
+	}
+	for name, spec := range cases {
+		if _, err := NormalizeWaitSpec(spec, "sample.tool"); err == nil {
+			t.Fatalf("%s: expected error, got nil", name)
+		}
+	}
+}
+
+func TestNormalizeWaitSpecNilReturnsNil(t *testing.T) {
+	out, err := NormalizeWaitSpec(nil, "sample.tool")
+	if err != nil || out != nil {
+		t.Fatalf("NormalizeWaitSpec(nil) = %#v, %v", out, err)
+	}
+}

internal/corecmd/contract_decl.go

@@ -42,6 +42,7 @@ type ContractDecl struct {
 	Positionals []contract.RuntimeSchemaPositional
 	Parameters  []contract.ParamDecl
 	DryRun      *contract.DryRunSpec
+	Wait        *contract.WaitSpec
 	Result      *contract.ResultSpec
 	Pagination  *contract.PaginationSpec
 	Interface   *contract.InterfaceSpec
@@ -146,6 +147,9 @@ func (s ContractDecl) empty() bool {
 	if s.DryRun != nil && strings.TrimSpace(s.DryRun.PreviewKind) != "" {
 		return false
 	}
+	if s.Wait != nil && strings.TrimSpace(s.Wait.Mode) != "" {
+		return false
+	}
 	if s.Result != nil {
 		return false
 	}

internal/corecmd/contractfinal/final_test.go

@@ -200,6 +200,13 @@ func TestFrameworkContractFinalDeepCopyAndSafetyConflicts(t *testing.T) {
 		Parameters:  []contract.ParamDecl{{Name: "mode", Enum: []string{"a"}, Required: boolPointer(true)}},
 		Safety:      &contract.SafetySpec{Effect: " read ", EffectSource: " source ", Risk: " low ", Confirmation: " not_required ", Idempotency: " idempotent "},
 		DryRun:      &contract.DryRunSpec{PreviewKind: "plan"},
+		Wait: &contract.WaitSpec{
+			Mode:          contract.WaitModePoll,
+			PollCommand:   "oa approval-instance get",
+			StatusQuery:   "result.status",
+			Terminal:      map[string]contract.ResultOutcome{"COMPLETED": contract.ResultOutcomeSuccess, "REJECTED": contract.ResultOutcomeFailure},
+			PendingValues: []string{"NEW"},
+		},
 		Result: &contract.ResultSpec{
 			Outcomes:   []contract.ResultOutcome{contract.ResultOutcomeSuccess},
 			DataSchema: []byte(`{"type":"object"}`), SensitivePaths: []string{"token"},
@@ -218,6 +225,8 @@ func TestFrameworkContractFinalDeepCopyAndSafetyConflicts(t *testing.T) {
 	if !ok || got.Result == payload.Result || got.Pagination == payload.Pagination || got.Interface == payload.Interface || got.Selection == payload.Selection || got.Identity == payload.Identity {
 		t.Fatalf("payload not deeply cloned: %#v", got)
 	}
+	payload.Wait.Terminal["COMPLETED"] = contract.ResultOutcomeFailure
+	payload.Wait.PendingValues[0] = "mutated"
 	payload.Parameters[0].Enum[0] = "changed"
 	*payload.Parameters[0].Required = false
 	*payload.Selection.ExampleDispositions[0].Index = 9
@@ -226,6 +235,10 @@ func TestFrameworkContractFinalDeepCopyAndSafetyConflicts(t *testing.T) {
 	if again.Parameters[0].Enum[0] != "a" || !*again.Parameters[0].Required || *again.Selection.ExampleDispositions[0].Index != 1 || !*again.Selection.Reviewed {
 		t.Fatalf("stored payload aliased input: %#v", again)
 	}
+	if again.Wait == payload.Wait || again.Wait.Terminal["COMPLETED"] != contract.ResultOutcomeSuccess || again.Wait.PendingValues[0] != "NEW" {
+		t.Fatalf("wait spec aliased input: %#v", again.Wait)
+		t.Fatalf("stored payload aliased input: %#v", again)
+	}
 
 	matching := &cobra.Command{Use: "matching"}
 	t.Cleanup(func() { ClearRuntimeContractFinalForTest(matching) })

internal/corecmd/contractfinal/store.go

@@ -74,6 +74,15 @@ func cloneContractFinalPayload(in contract.ContractFinalPayload) contract.Contra
 		value := *in.DryRun
 		out.DryRun = &value
 	}
+	if in.Wait != nil {
+		value := *in.Wait
+		value.Terminal = make(map[string]contract.ResultOutcome, len(in.Wait.Terminal))
+		for status, outcome := range in.Wait.Terminal {
+			value.Terminal[status] = outcome
+		}
+		value.PendingValues = cloneSlice(in.Wait.PendingValues)
+		out.Wait = &value
+	}
 	if in.Result != nil {
 		value := *in.Result
 		value.Outcomes = cloneSlice(in.Result.Outcomes)

internal/corecmd/corecmd.go

@@ -49,12 +49,16 @@ package corecmd
 
 import (
 	"bufio"
+	"context"
+	"encoding/json"
 	"errors"
 	"fmt"
 	"io"
+	"math"
 	"os"
 	"strconv"
 	"strings"
+	"time"
 
 	"github.com/mattn/go-isatty"
 	"github.com/spf13/cobra"
@@ -64,6 +68,7 @@ import (
 	"github.com/DingTalk-Real-AI/dingtalk-workspace-cli/internal/corecmd/runtimeannotate"
 	apperrors "github.com/DingTalk-Real-AI/dingtalk-workspace-cli/internal/errors"
 	"github.com/DingTalk-Real-AI/dingtalk-workspace-cli/internal/output"
+	"github.com/DingTalk-Real-AI/dingtalk-workspace-cli/internal/wait"
 	"github.com/DingTalk-Real-AI/dingtalk-workspace-cli/pkg/cmdutil"
 )
 
@@ -145,6 +150,15 @@ type FlagSpec struct {
 	// alike) and makes a whitespace-only value count as empty in required checks.
 	Trim bool
 
+	// Input declares extra input sources for a KindString flag beyond the
+	// literal command-line value: InputFile enables @path (value replaced by
+	// the file content), InputStdin enables - (value replaced by stdin).
+	// "@@value" always escapes to the literal "@value". Only explicit CLI
+	// tokens are resolved; EnvVar fallback and registration defaults pass
+	// through unchanged. Resolution runs before required/enum/constraint/
+	// Validate checks, so they see the payload content. Empty = flag value only.
+	Input []string
+
 	// Schema parameter final facts (embedded to dws.schema.*; assembly pass-through).
 	Enum              []string // accepted values
 	Format            string   // machine-readable format (e.g. uri)
@@ -273,6 +287,21 @@ type Spec struct {
 	// Orchestrate executes a multi-step command; it assembles whatever payloads
 	// it needs from the Ctx.
 	Orchestrate func(c *Ctx) error
+	// WaitPoll executes one poll of the declared Contract.Wait capability.
+	// Exactly one poll is one call; cadence, status extraction, and outcome
+	// mapping belong to the framework wait phase. Required for poll/auto
+	// declarations — a declared capability without a runtime implementation
+	// can never honor --wait, so New rejects the pairing at construction.
+	// ctx is the wait-phase deadline (--wait-timeout); leaf I/O must honor
+	// it so a blocked poll cannot outlive the declared timeout.
+	WaitPoll func(ctx context.Context, c *Ctx) (wait.PollDoc, error)
+	// WaitEvents opens the push subscription of the declared Contract.Wait
+	// capability (event/auto modes). The framework owns correlation and
+	// status mapping; the leaf owns the transport. Auto mode falls back to
+	// WaitPoll when the stream ends before a terminal status. ctx is the
+	// same wait-phase deadline as WaitPoll; subscription setup must honor
+	// it so --wait-timeout can cancel a blocked subscribe.
+	WaitEvents func(ctx context.Context, c *Ctx) (wait.EventStream, error)
 }
 
 // Ctx is the framework-neutral execution context handed to Invoke/Orchestrate.
@@ -346,6 +375,15 @@ func (c *Ctx) Changed(name string) bool { return c.cmd.Flags().Changed(name) }
 // DryRun reports the effective global --dry-run.
 func (c *Ctx) DryRun() bool { return BoolFlag(c.cmd, "dry-run") }
 
+// Wait reports the effective --wait flag. It is false on commands that did
+// not declare the capability: the flag is not registered there, so passing it
+// is an unknown-flag error rather than a silently ignored value.
+func (c *Ctx) Wait() bool { return BoolFlag(c.cmd, waitFlagName) }
+
+// WaitTimeoutSecs reports the effective --wait-timeout in seconds (flag
+// value, then the declared default, then the framework default).
+func (c *Ctx) WaitTimeoutSecs() int { return waitTimeoutSecs(c.cmd) }
+
 // Yes reports the effective global --yes.
 func (c *Ctx) Yes() bool { return BoolFlag(c.cmd, "yes") }
 
@@ -365,6 +403,9 @@ func New(spec Spec) *cobra.Command {
 	validateDispatchDecl(spec)
 	validateSafetySpec(spec)
 	validateContractDecl(spec)
+	normalizeWaitDecl(&spec)
+	validateWaitDecl(spec)
+	validateInputSpecs(spec.Use, spec.Flags)
 	// Help prose inherits the declaration when not authored separately:
 	// Selection.Examples (already contract-validated against the real flags)
 	// double as the --help Example block, keeping one authored source.
@@ -380,6 +421,7 @@ func New(spec Spec) *cobra.Command {
 		Hidden:  spec.Hidden,
 	}
 	RegisterFlags(cmd, spec.Flags)
+	registerWaitFlags(cmd, spec)
 	ValidateConstraintDecls(spec.Use, spec.Flags, spec.Constraints)
 	embedContractIntoSchema(cmd, spec)
 	AnnotateConstraints(cmd, spec.Constraints)
@@ -445,6 +487,10 @@ func New(spec Spec) *cobra.Command {
 			if err != nil {
 				return err
 			}
+			result, err = runDeclaredWaitPhase(cmd, args, spec, result)
+			if err != nil {
+				return err
+			}
 			return output.StoreResult(cmd.Context(), result)
 		}
 		return spec.Invoke(ctx, toolArgs)
@@ -476,6 +522,11 @@ func runDeclaredPreflight(cmd *cobra.Command, args []string, spec Spec) error {
 			return err
 		}
 	}
+	// Input resolution rewrites explicit @file / stdin values in place so the
+	// required/enum/constraint/Validate stages below check the payload content.
+	if err := resolveInputFlags(cmd, spec.Flags); err != nil {
+		return err
+	}
 	if err := ValidateRequired(cmd, spec.Flags); err != nil {
 		return err
 	}
@@ -491,6 +542,293 @@ func runDeclaredPreflight(cmd *cobra.Command, args []string, spec Spec) error {
 	return nil
 }
 
+// Wait-phase framework flags. They are registered natively on the leaf (never
+// as FlagSpec) so they cannot leak into toolArgs / MCP payloads: --wait is a
+// client-side execution modifier, not a backend parameter. On commands that
+// did not declare Contract.Wait the flags do not exist, so passing --wait
+// fails as an unknown flag instead of being silently ignored.
+const (
+	waitFlagName        = "wait"
+	waitTimeoutFlagName = "wait-timeout"
+	defaultWaitTimeoutS = 300
+)
+
+// DefaultWaitTimeoutSecs is the framework default for --wait-timeout when the
+// declaration carries no reviewed default.
+const DefaultWaitTimeoutSecs = defaultWaitTimeoutS
+
+// normalizeWaitDecl rewrites the spec's declared Wait in place with its
+// canonical form (contract.NormalizeWaitSpec): trimmed status values,
+// duplicate/conflict rejection, defensive copy. The runtime wait phase and
+// AttachContract both read spec.Contract.Wait, so normalizing once at
+// construction guarantees the wait engine, the Schema wire, and the
+// registered ContractFinal payload all see identical status tables — a
+// padded declaration can no longer publish a Schema its own runtime treats
+// as unknown statuses. An invalid declaration panics here, next to the
+// authoring mistake, with the same message Validate reports.
+func normalizeWaitDecl(spec *Spec) {
+	decl := spec.Contract.Wait
+	if decl == nil || strings.TrimSpace(decl.Mode) == "" {
+		return
+	}
+	normalized, err := contract.NormalizeWaitSpec(decl, spec.Contract.Identity.CanonicalPath)
+	if err != nil {
+		panic(fmt.Sprintf("command %q has invalid Contract.Wait: %v", spec.Use, err))
+	}
+	spec.Contract.Wait = normalized
+}
+
+// validateWaitDecl enforces the declaration ⇄ implementation pairing at build
+// time: a declared Contract.Wait without a WaitPoll hook is a capability the
+// command can never honor, and a WaitPoll hook without the declaration has no
+// flags or Schema capability to serve. The declaration also requires the
+// ResultInvoke dispatcher: only the unified-result envelope can be closed
+// into the terminal outcome (error.type "wait", exit code 8) and the timed-out
+// pending form — legacy Invoke/Orchestrate/RunE paths emit their own output
+// and would observe a failure terminal while still exiting 0. All three
+// mismatches are programming errors.
+func validateWaitDecl(spec Spec) {
+	decl := spec.Contract.Wait
+	declared := decl != nil && strings.TrimSpace(decl.Mode) != ""
+	if !declared {
+		if spec.WaitPoll != nil || spec.WaitEvents != nil {
+			panic(fmt.Sprintf(
+				"command %q sets a wait hook without declaring Contract.Wait: the wait flags and Schema capability come from the declaration",
+				spec.Use))
+		}
+		return
+	}
+	if spec.ResultInvoke == nil {
+		panic(fmt.Sprintf(
+			"command %q declares Contract.Wait without ResultInvoke: wait closes the unified-result envelope, which legacy Invoke/Orchestrate/RunE paths cannot rewrite",
+			spec.Use))
+	}
+	mode := strings.TrimSpace(decl.Mode)
+	needsPoll := mode == contract.WaitModePoll || mode == contract.WaitModeAuto
+	needsEvent := mode == contract.WaitModeEvent || mode == contract.WaitModeAuto
+	if needsPoll && spec.WaitPoll == nil {
+		panic(fmt.Sprintf(
+			"command %q declares wait mode %s but sets no WaitPoll: a declared wait capability must carry its runtime poll implementation",
+			spec.Use, mode))
+	}
+	if needsEvent && spec.WaitEvents == nil {
+		panic(fmt.Sprintf(
+			"command %q declares wait mode %s but sets no WaitEvents: a declared event wait must carry its runtime subscription",
+			spec.Use, mode))
+	}
+	if !needsPoll && spec.WaitPoll != nil {
+		panic(fmt.Sprintf(
+			"command %q declares wait mode %s but sets WaitPoll: the declaration decides which hooks run",
+			spec.Use, mode))
+	}
+	if !needsEvent && spec.WaitEvents != nil {
+		panic(fmt.Sprintf(
+			"command %q declares wait mode %s but sets WaitEvents: the declaration decides which hooks run",
+			spec.Use, mode))
+	}
+}
+
+// registerWaitFlags adds --wait / --wait-timeout to a leaf that declared
+// Contract.Wait. The timeout default is the reviewed declaration, falling
+// back to DefaultWaitTimeoutSecs.
+func registerWaitFlags(cmd *cobra.Command, spec Spec) {
+	decl := spec.Contract.Wait
+	if decl == nil || strings.TrimSpace(decl.Mode) == "" {
+		return
+	}
+	cmd.Flags().Bool(waitFlagName, false,
+		"等待到达命令声明的终态（如审批完成、导出结束）后再返回；未声明该能力的命令不接受此 flag")
+	timeoutDefault := decl.DefaultTimeoutSecs
+	if timeoutDefault <= 0 {
+		timeoutDefault = DefaultWaitTimeoutSecs
+	}
+	cmd.Flags().Int(waitTimeoutFlagName, timeoutDefault,
+		"等待超时秒数；超时以 pending 结束（异步受理不是失败）")
+}
+
+// runDeclaredWaitPhase runs the declared wait loop after a successful
+// ResultInvoke dispatch and closes the accepted unified envelope into the
+// wait outcome (validateWaitDecl guarantees the ResultInvoke pairing).
+// Only a pending accepted result is waitable: success / failure / partial
+// are already terminal and must be returned unchanged. Waiting on a
+// business failure would let WithOutcome(..., success) overwrite it into
+// an illegal success-with-error envelope.
+func runDeclaredWaitPhase(cmd *cobra.Command, args []string, spec Spec, result output.CommandResult) (output.CommandResult, error) {
+	if !BoolFlag(cmd, waitFlagName) {
+		return result, nil
+	}
+	if result == nil || result.Outcome() != output.OutcomePending {
+		return result, nil
+	}
+	decl := spec.Contract.Wait
+	timeout, err := waitTimeoutDuration(int64(waitTimeoutSecs(cmd)))
+	if err != nil {
+		return result, err
+	}
+	ctx := newCtx(cmd, args, spec.Flags)
+	outcome, err := runWaitLoop(cmd.Context(), decl, timeout, spec, ctx, result)
+	if err != nil {
+		return result, err
+	}
+	if outcome.TimedOut {
+		cmd.PrintErrf("等待超时（%s）：当前状态 %q，未到达终态，以 pending 结束\n", timeout, outcome.Status)
+		return output.WithOutcome(result, output.OutcomePending,
+			output.WithOperationTimedOut(outcome.Status)), nil
+	}
+	if outcome.Outcome == contract.ResultOutcomeFailure {
+		return output.WithOutcome(result, output.OutcomeFailure,
+			output.WithOperationTerminalState(outcome.Status),
+			output.WithErrorInfo(&output.ErrorInfo{
+				Type:    "wait",
+				Subtype: "terminal_failure",
+				Message: fmt.Sprintf("等待到达失败终态：%s", outcome.Status),
+			})), nil
+	}
+	return output.WithOutcome(result, output.OutcomeSuccess,
+		output.WithOperationTerminalState(outcome.Status)), nil
+}
+
+// runWaitLoop executes the declared wait mode. One deadline spans the event
+// phase and an auto-mode poll fallback (the inner loops run without their
+// own timeouts and inherit this context's deadline). The deadline is
+// forwarded to WaitPoll / WaitEvents and bound onto the cobra command so
+// leaf I/O that reads either the hook ctx or Command().Context() is
+// cancelled when --wait-timeout expires.
+func runWaitLoop(parent context.Context, decl *contract.WaitSpec, timeout time.Duration, spec Spec, ctx *Ctx, result output.CommandResult) (wait.Outcome, error) {
+	loopCtx := parent
+	if timeout > 0 {
+		var cancel context.CancelFunc
+		loopCtx, cancel = context.WithTimeout(parent, timeout)
+		defer cancel()
+	}
+	if ctx != nil && ctx.cmd != nil {
+		prev := ctx.cmd.Context()
+		ctx.cmd.SetContext(loopCtx)
+		defer ctx.cmd.SetContext(prev)
+	}
+	mode := strings.TrimSpace(decl.Mode)
+	if mode == contract.WaitModePoll {
+		return wait.Run(loopCtx, wait.LoopSpec{
+			StatusQuery: decl.StatusQuery,
+			Terminal:    decl.Terminal,
+			Pending:     decl.PendingValues,
+		}, func(pollCtx context.Context) (wait.PollDoc, error) {
+			return spec.WaitPoll(pollCtx, ctx)
+		})
+	}
+	resource, err := waitResource(decl, result)
+	if err != nil {
+		return wait.Outcome{}, err
+	}
+	stream, err := spec.WaitEvents(loopCtx, ctx)
+	if err != nil {
+		if loopCtx.Err() != nil {
+			// Subscribe blocked until the wait deadline: same contract as a
+			// cancelled poll — close as timed-out pending, do not surface
+			// ctx.Err() as a subscription failure (and do not poll-fallback
+			// in auto mode; the shared deadline is already exhausted).
+			return wait.Outcome{Outcome: contract.ResultOutcomePending, TimedOut: true}, nil
+		}
+		if mode == contract.WaitModeAuto {
+			// Subscription failed before any event: fall back to polling.
+			return pollWithSpec(loopCtx, decl, spec, ctx)
+		}
+		return wait.Outcome{}, fmt.Errorf("wait: event subscription failed: %w", err)
+	}
+	eventSpec := wait.EventLoopSpec{
+		StatusQuery: decl.StatusQuery,
+		MatchField:  decl.MatchField,
+		Terminal:    decl.Terminal,
+		Pending:     decl.PendingValues,
+	}
+	outcome, err := wait.RunEvent(loopCtx, eventSpec, resource, stream)
+	if err == nil {
+		return outcome, nil
+	}
+	if mode == contract.WaitModeAuto && errors.Is(err, wait.ErrEventStreamEnded) {
+		// Stream ended before a terminal status: fall back to polling under
+		// the same deadline.
+		return pollWithSpec(loopCtx, decl, spec, ctx)
+	}
+	return outcome, err
+}
+
+// pollWithSpec runs the poll loop for an auto-mode fallback.
+func pollWithSpec(loopCtx context.Context, decl *contract.WaitSpec, spec Spec, ctx *Ctx) (wait.Outcome, error) {
+	return wait.Run(loopCtx, wait.LoopSpec{
+		StatusQuery: decl.StatusQuery,
+		Terminal:    decl.Terminal,
+		Pending:     decl.PendingValues,
+	}, func(pollCtx context.Context) (wait.PollDoc, error) {
+		return spec.WaitPoll(pollCtx, ctx)
+	})
+}
+
+// waitResource resolves the resource identifier an event stream correlates
+// against, from the accepted result data via the declared ResourceQuery.
+// result.Data() returns any deep-copied business payload, which may be a
+// map[string]any, struct, or struct pointer. We normalize via JSON round-trip
+// to support all valid result types uniformly.
+func waitResource(decl *contract.WaitSpec, result output.CommandResult) (string, error) {
+	raw := result.Data()
+	if raw == nil {
+		return "", fmt.Errorf("wait: accepted result data is nil; cannot resolve resource %q", decl.ResourceQuery)
+	}
+	// Fast path: already a map.
+	if data, ok := raw.(map[string]any); ok {
+		resource, ok := wait.ExtractStatus(wait.PollDoc(data), decl.ResourceQuery)
+		if !ok || strings.TrimSpace(resource) == "" {
+			return "", fmt.Errorf("wait: resource query %q not found in accepted result data", decl.ResourceQuery)
+		}
+		return resource, nil
+	}
+	// Slow path: struct or struct pointer. Normalize via JSON round-trip.
+	jsonBytes, err := json.Marshal(raw)
+	if err != nil {
+		return "", fmt.Errorf("wait: accepted result data cannot be serialized to JSON: %w", err)
+	}
+	var data map[string]any
+	if err := json.Unmarshal(jsonBytes, &data); err != nil {
+		return "", fmt.Errorf("wait: accepted result data is not an object; cannot resolve resource %q", decl.ResourceQuery)
+	}
+	resource, ok := wait.ExtractStatus(wait.PollDoc(data), decl.ResourceQuery)
+	if !ok || strings.TrimSpace(resource) == "" {
+		return "", fmt.Errorf("wait: resource query %q not found in accepted result data", decl.ResourceQuery)
+	}
+	return resource, nil
+}
+
+// waitTimeoutSecs resolves the effective timeout. The flag is registered
+// with the reviewed declaration default (or the framework default), so the
+// flag value is authoritative; a non-positive explicit value falls back to
+// the framework default.
+func waitTimeoutSecs(cmd *cobra.Command) int {
+	if value, err := cmd.Flags().GetInt(waitTimeoutFlagName); err == nil && value > 0 {
+		return value
+	}
+	return DefaultWaitTimeoutSecs
+}
+
+// maxWaitTimeoutSecs is the largest second count that still fits in a
+// time.Duration. Multiplying a larger int by time.Second overflows to a
+// non-positive duration, which would skip the deadline and wait forever.
+const maxWaitTimeoutSecs = math.MaxInt64 / int64(time.Second)
+
+// waitTimeoutDuration converts a resolved second count into the wait-phase
+// deadline. Values that cannot be represented as a positive time.Duration
+// are rejected as validation errors instead of silently disabling timeout.
+func waitTimeoutDuration(secs int64) (time.Duration, error) {
+	if secs <= 0 {
+		secs = DefaultWaitTimeoutSecs
+	}
+	if secs > maxWaitTimeoutSecs {
+		return 0, apperrors.NewValidation(fmt.Sprintf(
+			"参数 --%s 取值 %d 超出可表示范围（最大 %d 秒）",
+			waitTimeoutFlagName, secs, maxWaitTimeoutSecs))
+	}
+	return time.Duration(secs) * time.Second, nil
+}
+
 // validateDispatchDecl enforces "exactly one dispatcher" at build time. Like
 // ValidateConstraintDecls this panics: a spec with no runnable body (or with two
 // competing ones) is a programming error that every test and startup path should
@@ -1393,6 +1731,13 @@ func AttachContract(cmd *cobra.Command, safety contract.SafetySpec, decl Contrac
 		d.PreviewKind = strings.TrimSpace(d.PreviewKind)
 		payload.DryRun = &d
 	}
+	if decl.Wait != nil && strings.TrimSpace(decl.Wait.Mode) != "" {
+		waitSpec, err := contract.NormalizeWaitSpec(decl.Wait, decl.Identity.CanonicalPath)
+		if err != nil {
+			panic(fmt.Sprintf("command %q has invalid Contract.Wait: %v", cmd.Name(), err))
+		}
+		payload.Wait = waitSpec
+	}
 	if decl.Result != nil {
 		result, err := contract.NormalizeResultSpec(decl.Result, decl.Identity.CanonicalPath)
 		if err != nil {

internal/corecmd/corecmd_test.go

@@ -26,6 +26,7 @@ import (
 	"github.com/DingTalk-Real-AI/dingtalk-workspace-cli/internal/corecmd/contractfinal"
 	"github.com/DingTalk-Real-AI/dingtalk-workspace-cli/internal/corecmd/runtimeannotate"
 	apperrors "github.com/DingTalk-Real-AI/dingtalk-workspace-cli/internal/errors"
+	"github.com/DingTalk-Real-AI/dingtalk-workspace-cli/internal/output"
 	"github.com/DingTalk-Real-AI/dingtalk-workspace-cli/internal/testseam"
 	"github.com/spf13/cobra"
 )
@@ -2120,3 +2121,111 @@ func TestCrossPlatformCoverageEmbedContractSkipsBlankAndHiddenFlags(t *testing.T
 		}
 	}
 }
+
+// TestWaitResourceStronglyTypedDTOs verifies waitResource handles struct and
+// struct pointer results via JSON normalization (P1 fix for auto-CR).
+func TestWaitResourceStronglyTypedDTOs(t *testing.T) {
+	type TaskDTO struct {
+		TaskID string `json:"task_id"`
+		Status string `json:"status"`
+	}
+	type NestedDTO struct {
+		Meta struct {
+			ResourceID string `json:"resource_id"`
+		} `json:"meta"`
+	}
+
+	tests := []struct {
+		name          string
+		data          any
+		query         string
+		wantResource  string
+		wantErrSubstr string
+	}{
+		{
+			name:         "map[string]any fast path",
+			data:         map[string]any{"task_id": "abc123"},
+			query:        "task_id",
+			wantResource: "abc123",
+		},
+		{
+			name:         "struct value",
+			data:         TaskDTO{TaskID: "struct-456", Status: "running"},
+			query:        "task_id",
+			wantResource: "struct-456",
+		},
+		{
+			name:         "struct pointer",
+			data:         &TaskDTO{TaskID: "ptr-789", Status: "pending"},
+			query:        "task_id",
+			wantResource: "ptr-789",
+		},
+		{
+			name:         "nested struct dotted query",
+			data:         NestedDTO{},
+			query:        "meta.resource_id",
+			wantResource: "",
+			wantErrSubstr: "not found",
+		},
+		{
+			name: "nested struct with value",
+			data: func() NestedDTO {
+				var d NestedDTO
+				d.Meta.ResourceID = "nested-xyz"
+				return d
+			}(),
+			query:        "meta.resource_id",
+			wantResource: "nested-xyz",
+		},
+		{
+			name:          "nil data",
+			data:          nil,
+			query:         "task_id",
+			wantErrSubstr: "nil",
+		},
+		{
+			name:          "non-object data (string)",
+			data:          "not-an-object",
+			query:         "task_id",
+			wantErrSubstr: "not an object",
+		},
+		{
+			name:          "non-object data (slice)",
+			data:          []string{"a", "b"},
+			query:         "task_id",
+			wantErrSubstr: "not an object",
+		},
+		{
+			name:          "missing query field in struct",
+			data:          TaskDTO{TaskID: "abc"},
+			query:         "nonexistent",
+			wantErrSubstr: "not found",
+		},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			decl := &contract.WaitSpec{
+				ResourceQuery: tt.query,
+			}
+			result := output.Success(tt.data)
+			resource, err := waitResource(decl, result)
+
+			if tt.wantErrSubstr != "" {
+				if err == nil {
+					t.Fatalf("expected error containing %q, got nil", tt.wantErrSubstr)
+				}
+				if !strings.Contains(err.Error(), tt.wantErrSubstr) {
+					t.Errorf("error = %q; want substring %q", err.Error(), tt.wantErrSubstr)
+				}
+				return
+			}
+			if err != nil {
+				t.Fatalf("unexpected error: %v", err)
+			}
+			if resource != tt.wantResource {
+				t.Errorf("resource = %q; want %q", resource, tt.wantResource)
+			}
+		})
+	}
+}

internal/corecmd/input.go

@@ -0,0 +1,191 @@
+// Copyright 2026 Alibaba Group
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+package corecmd
+
+import (
+	"fmt"
+	"io"
+	"os"
+	"strings"
+
+	"github.com/spf13/cobra"
+
+	apperrors "github.com/DingTalk-Real-AI/dingtalk-workspace-cli/internal/errors"
+	"github.com/DingTalk-Real-AI/dingtalk-workspace-cli/pkg/cmdutil"
+)
+
+// FlagSpec.Input source constants. They declare extra input sources for a
+// KindString flag beyond the literal command-line value, mirroring the
+// lark-cli Flag.Input capability so large payloads (markdown, JSON, CSV)
+// never need shell quoting.
+const (
+	// InputFile allows the flag value @path to be replaced by the file content.
+	InputFile = "file"
+	// InputStdin allows the flag value - to be replaced by the stdin content.
+	InputStdin = "stdin"
+)
+
+// utf8BOM is stripped from file/stdin content so a Windows-edited payload
+// cannot corrupt the first CSV cell or break JSON parsing downstream.
+const utf8BOM = "\ufeff"
+
+// validateInputSpecs rejects malformed Input declarations at build time. Like
+// the other declaration checks this panics: a bad Input spec is a programming
+// error every test and startup path should trip immediately.
+func validateInputSpecs(use string, flags []FlagSpec) {
+	for _, flag := range flags {
+		if len(flag.Input) == 0 {
+			continue
+		}
+		if flag.Kind != KindString {
+			panic(fmt.Sprintf(
+				"command %q flag %q: Input is only supported on KindString flags",
+				use, flag.Name))
+		}
+		seen := map[string]bool{}
+		for _, source := range flag.Input {
+			if source != InputFile && source != InputStdin {
+				panic(fmt.Sprintf(
+					"command %q flag %q: unknown Input source %q (allowed: %s, %s)",
+					use, flag.Name, source, InputFile, InputStdin))
+			}
+			if seen[source] {
+				panic(fmt.Sprintf(
+					"command %q flag %q: duplicate Input source %q",
+					use, flag.Name, source))
+			}
+			seen[source] = true
+		}
+	}
+}
+
+// inputSupports reports whether the flag declared the given input source.
+func inputSupports(flag FlagSpec, source string) bool {
+	for _, s := range flag.Input {
+		if s == source {
+			return true
+		}
+	}
+	return false
+}
+
+// explicitInputFlagName picks the declared name that carries the user's
+// explicit token, mirroring rawValue's order and usability judgement exactly:
+// the main flag wins only when changed and usable (trim-judged only when
+// Trim is set), then declared aliases in order. Diverging here could rewrite
+// an alias that rawValue would shadow (or vice versa). EnvVar fallback and
+// registration defaults are never input-resolved — @file only applies to what
+// the user literally typed.
+func explicitInputFlagName(cmd *cobra.Command, flag FlagSpec) string {
+	usable := func(v string) bool {
+		if flag.Trim {
+			v = strings.TrimSpace(v)
+		}
+		return v != ""
+	}
+	for _, name := range append([]string{flag.Name}, flag.Aliases...) {
+		if cmd.Flags().Changed(name) && usable(cmdutil.MustGetFlag(cmd, name)) {
+			return name
+		}
+	}
+	return ""
+}
+
+// resolveInputFlags rewrites the explicit values of Input-declaring flags
+// before required/enum/constraint/Validate checks, so every downstream stage
+// sees the real payload content. Semantics:
+//
+//   - "-" reads stdin; a process has a single stdin, so a second Input flag
+//     using "-" in the same invocation is rejected.
+//   - "@path" reads the named file (leading/trailing whitespace trimmed).
+//   - "@@value" escapes to the literal "@value" and is not source-resolved.
+//
+// Resolution runs once per invocation inside runDeclaredPreflight. It
+// rewrites the cobra flag value in place (Set keeps Changed=true), so
+// fallback-chain readers (EffectiveValue/BuildArgs) observe the content.
+//
+// Interaction with confirmation: stdin is consumed here, before the Safety
+// prompt of a user_required write would read it; such an interactive prompt
+// then sees EOF and fails closed with confirmation_required. Write commands
+// declaring InputStdin must be invoked with --yes (or --dry-run).
+func resolveInputFlags(cmd *cobra.Command, flags []FlagSpec) error {
+	stdinConsumed := false
+	for _, flag := range flags {
+		if len(flag.Input) == 0 {
+			continue
+		}
+		name := explicitInputFlagName(cmd, flag)
+		if name == "" {
+			continue
+		}
+		raw := cmdutil.MustGetFlag(cmd, name)
+		// Trim flags judge usability on the trimmed value (rawValue), so the
+		// prefix check must agree or " @path" would slip through unresolved.
+		if flag.Trim {
+			raw = strings.TrimSpace(raw)
+		}
+
+		switch {
+		case raw == "-":
+			if !inputSupports(flag, InputStdin) {
+				return apperrors.NewValidation(
+					fmt.Sprintf("参数 --%s 不支持 stdin 输入（-）", flag.Name))
+			}
+			if stdinConsumed {
+				return apperrors.NewValidation(
+					fmt.Sprintf("参数 --%s：stdin（-）只能被一个参数使用", flag.Name),
+					apperrors.WithHint(fmt.Sprintf(
+						"一个进程只有一份 stdin，其余参数请内联传值或用 @文件路径（如 --%s @./payload.json）",
+						flag.Name)))
+			}
+			stdinConsumed = true
+			data, err := io.ReadAll(cmd.InOrStdin())
+			if err != nil {
+				return apperrors.NewValidation(
+					fmt.Sprintf("参数 --%s 读取 stdin 失败：%v", flag.Name, err))
+			}
+			// Setting a registered string flag cannot fail.
+			_ = cmd.Flags().Set(name, strings.TrimPrefix(string(data), utf8BOM))
+
+		case strings.HasPrefix(raw, "@@"):
+			// Escape: strip the first @, keep the rest as a literal inline value.
+			_ = cmd.Flags().Set(name, raw[1:])
+
+		case strings.HasPrefix(raw, "@"):
+			if !inputSupports(flag, InputFile) {
+				return apperrors.NewValidation(
+					fmt.Sprintf("参数 --%s 不支持文件输入（@路径）", flag.Name))
+			}
+			path := strings.TrimSpace(raw[1:])
+			if path == "" {
+				return apperrors.NewValidation(
+					fmt.Sprintf("参数 --%s：@ 后的文件路径不能为空", flag.Name))
+			}
+			data, err := os.ReadFile(path)
+			if err != nil {
+				var opts []apperrors.Option
+				if inputSupports(flag, InputStdin) {
+					// Rejected @file paths are usually absolute (temp files).
+					// Steer toward stdin rather than copying the file around.
+					opts = append(opts, apperrors.WithHint(fmt.Sprintf(
+						"该参数也支持 stdin：把文件内容管道进命令并传 --%s -", flag.Name)))
+				}
+				return apperrors.NewValidation(
+					fmt.Sprintf("参数 --%s 读取文件 %q 失败：%v", flag.Name, path, err), opts...)
+			}
+			_ = cmd.Flags().Set(name, strings.TrimPrefix(string(data), utf8BOM))
+		}
+	}
+	return nil
+}

internal/corecmd/input_test.go

@@ -0,0 +1,334 @@
+// Copyright 2026 Alibaba Group
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+package corecmd
+
+import (
+	"os"
+	"path/filepath"
+	"strings"
+	"testing"
+
+	"github.com/spf13/cobra"
+)
+
+// newInputCommand builds a Spec whose Invoke captures toolArgs, so tests can
+// assert what the full pipeline (resolution → validation → BuildArgs) ships.
+func newInputCommand(flags []FlagSpec, captured *map[string]any) *cobra.Command {
+	return New(Spec{
+		Use:   "t",
+		Short: "t",
+		Flags: flags,
+		Invoke: func(c *Ctx, toolArgs map[string]any) error {
+			*captured = toolArgs
+			return nil
+		},
+	})
+}
+
+func runInputCommand(t *testing.T, cmd *cobra.Command, args ...string) error {
+	t.Helper()
+	cmd.SetArgs(args)
+	return cmd.Execute()
+}
+
+func writeInputFile(t *testing.T, content string) string {
+	t.Helper()
+	path := filepath.Join(t.TempDir(), "payload.txt")
+	if err := os.WriteFile(path, []byte(content), 0o600); err != nil {
+		t.Fatalf("write fixture: %v", err)
+	}
+	return path
+}
+
+func TestCrossPlatformCoverageResolveInputFlagsFile(t *testing.T) {
+	var got map[string]any
+	cmd := newInputCommand([]FlagSpec{
+		{Name: "content", Usage: "C", Bind: "content", Input: []string{InputFile, InputStdin}},
+	}, &got)
+	path := writeInputFile(t, "# hello\nworld")
+
+	if err := runInputCommand(t, cmd, "--content", "@"+path); err != nil {
+		t.Fatalf("execute: %v", err)
+	}
+	if got["content"] != "# hello\nworld" {
+		t.Fatalf("toolArgs[content] = %q", got["content"])
+	}
+}
+
+func TestCrossPlatformCoverageResolveInputFlagsStdin(t *testing.T) {
+	var got map[string]any
+	cmd := newInputCommand([]FlagSpec{
+		{Name: "content", Usage: "C", Bind: "content", Input: []string{InputStdin}},
+	}, &got)
+	cmd.SetIn(strings.NewReader("piped payload"))
+
+	if err := runInputCommand(t, cmd, "--content", "-"); err != nil {
+		t.Fatalf("execute: %v", err)
+	}
+	if got["content"] != "piped payload" {
+		t.Fatalf("toolArgs[content] = %q", got["content"])
+	}
+}
+
+// The @@ escape keeps a literal leading @ inline and must not be treated as a
+// source reference even when Input is declared.
+func TestCrossPlatformCoverageResolveInputFlagsEscapedAtStaysInline(t *testing.T) {
+	var got map[string]any
+	cmd := newInputCommand([]FlagSpec{
+		{Name: "content", Usage: "C", Bind: "content", Input: []string{InputFile, InputStdin}},
+	}, &got)
+
+	if err := runInputCommand(t, cmd, "--content", "@@literal@x"); err != nil {
+		t.Fatalf("execute: %v", err)
+	}
+	if got["content"] != "@literal@x" {
+		t.Fatalf("toolArgs[content] = %q", got["content"])
+	}
+}
+
+func TestCrossPlatformCoverageResolveInputFlagsBOMStripped(t *testing.T) {
+	var got map[string]any
+	cmd := newInputCommand([]FlagSpec{
+		{Name: "content", Usage: "C", Bind: "content", Input: []string{InputFile}},
+	}, &got)
+	path := writeInputFile(t, "\ufeff{\"a\":1}")
+
+	if err := runInputCommand(t, cmd, "--content", "@"+path); err != nil {
+		t.Fatalf("execute: %v", err)
+	}
+	if got["content"] != "{\"a\":1}" {
+		t.Fatalf("toolArgs[content] = %q", got["content"])
+	}
+}
+
+// Required must be satisfied by the resolved payload, proving resolution runs
+// before the required stage.
+func TestCrossPlatformCoverageResolveInputFlagsSatisfiesRequired(t *testing.T) {
+	var got map[string]any
+	cmd := newInputCommand([]FlagSpec{
+		{Name: "content", Usage: "C", Bind: "content", Required: true, Input: []string{InputFile}},
+	}, &got)
+	path := writeInputFile(t, "payload")
+
+	if err := runInputCommand(t, cmd, "--content", "@"+path); err != nil {
+		t.Fatalf("execute: %v", err)
+	}
+	if got["content"] != "payload" {
+		t.Fatalf("toolArgs[content] = %q", got["content"])
+	}
+}
+
+// Enum validation sees the resolved content, not the "@path" token.
+func TestCrossPlatformCoverageResolveInputFlagsEnumValidatesResolvedContent(t *testing.T) {
+	var got map[string]any
+	cmd := newInputCommand([]FlagSpec{
+		{Name: "mode", Usage: "M", Bind: "mode", Enum: []string{"asc", "desc"}, Input: []string{InputFile}},
+	}, &got)
+	path := writeInputFile(t, "sideways")
+
+	err := runInputCommand(t, cmd, "--mode", "@"+path)
+	if err == nil || !strings.Contains(err.Error(), "不合法") {
+		t.Fatalf("expected enum rejection on resolved content, got %v", err)
+	}
+}
+
+// A value passed through a declared alias is resolved exactly like the main
+// name (fallback-chain parity).
+func TestCrossPlatformCoverageResolveInputFlagsAliasResolved(t *testing.T) {
+	var got map[string]any
+	cmd := newInputCommand([]FlagSpec{
+		{Name: "content", Usage: "C", Bind: "content", Aliases: []string{"body"}, Input: []string{InputFile}},
+	}, &got)
+	path := writeInputFile(t, "via alias")
+
+	if err := runInputCommand(t, cmd, "--body", "@"+path); err != nil {
+		t.Fatalf("execute: %v", err)
+	}
+	if got["content"] != "via alias" {
+		t.Fatalf("toolArgs[content] = %q", got["content"])
+	}
+}
+
+// When the main flag shadows a changed alias (rawValue usability order), the
+// shadowed alias must not be input-resolved: resolution targets exactly the
+// name the fallback chain will read. The whitespace main value is usable for
+// a non-Trim flag, and the alias path does not exist on purpose — a resolver
+// that wrongly picked the alias would fail the read instead of shipping "   ".
+func TestCrossPlatformCoverageResolveInputFlagsShadowedAliasNotResolved(t *testing.T) {
+	var got map[string]any
+	cmd := newInputCommand([]FlagSpec{
+		{Name: "content", Usage: "C", Bind: "content", Aliases: []string{"body"}, Input: []string{InputFile}},
+	}, &got)
+	missing := filepath.Join(t.TempDir(), "missing.txt")
+
+	if err := runInputCommand(t, cmd, "--content", "   ", "--body", "@"+missing); err != nil {
+		t.Fatalf("execute: %v", err)
+	}
+	if got["content"] != "   " {
+		t.Fatalf("toolArgs[content] = %q", got["content"])
+	}
+}
+
+func TestCrossPlatformCoverageResolveInputFlagsFileNotSupported(t *testing.T) {
+	var got map[string]any
+	cmd := newInputCommand([]FlagSpec{
+		{Name: "content", Usage: "C", Bind: "content", Input: []string{InputStdin}},
+	}, &got)
+
+	err := runInputCommand(t, cmd, "--content", "@/tmp/whatever.txt")
+	if err == nil || !strings.Contains(err.Error(), "不支持文件输入") {
+		t.Fatalf("expected file-input rejection, got %v", err)
+	}
+}
+
+func TestCrossPlatformCoverageResolveInputFlagsStdinNotSupported(t *testing.T) {
+	var got map[string]any
+	cmd := newInputCommand([]FlagSpec{
+		{Name: "content", Usage: "C", Bind: "content", Input: []string{InputFile}},
+	}, &got)
+	cmd.SetIn(strings.NewReader("x"))
+
+	err := runInputCommand(t, cmd, "--content", "-")
+	if err == nil || !strings.Contains(err.Error(), "不支持 stdin") {
+		t.Fatalf("expected stdin rejection, got %v", err)
+	}
+}
+
+func TestCrossPlatformCoverageResolveInputFlagsSingleStdinConsumer(t *testing.T) {
+	var got map[string]any
+	cmd := newInputCommand([]FlagSpec{
+		{Name: "first", Usage: "F", Bind: "first", Input: []string{InputStdin}},
+		{Name: "second", Usage: "S", Bind: "second", Input: []string{InputStdin}},
+	}, &got)
+	cmd.SetIn(strings.NewReader("x"))
+
+	err := runInputCommand(t, cmd, "--first", "-", "--second", "-")
+	if err == nil || !strings.Contains(err.Error(), "只能被一个参数使用") {
+		t.Fatalf("expected single-stdin rejection, got %v", err)
+	}
+}
+
+// failingReader (corecmd_test.go) fails every read, making the stdin error
+// branch reachable.
+func TestCrossPlatformCoverageResolveInputFlagsStdinReadFailure(t *testing.T) {
+	var got map[string]any
+	cmd := newInputCommand([]FlagSpec{
+		{Name: "content", Usage: "C", Bind: "content", Input: []string{InputStdin}},
+	}, &got)
+	cmd.SetIn(failingReader{})
+
+	err := runInputCommand(t, cmd, "--content", "-")
+	if err == nil || !strings.Contains(err.Error(), "读取 stdin 失败") {
+		t.Fatalf("expected stdin read failure, got %v", err)
+	}
+}
+
+func TestCrossPlatformCoverageResolveInputFlagsFileNotFound(t *testing.T) {
+	var got map[string]any
+	cmd := newInputCommand([]FlagSpec{
+		{Name: "content", Usage: "C", Bind: "content", Input: []string{InputFile, InputStdin}},
+	}, &got)
+
+	err := runInputCommand(t, cmd, "--content", "@"+filepath.Join(t.TempDir(), "missing.txt"))
+	if err == nil || !strings.Contains(err.Error(), "读取文件") {
+		t.Fatalf("expected read failure, got %v", err)
+	}
+}
+
+func TestCrossPlatformCoverageResolveInputFlagsEmptyPathAfterAt(t *testing.T) {
+	var got map[string]any
+	cmd := newInputCommand([]FlagSpec{
+		{Name: "content", Usage: "C", Bind: "content", Input: []string{InputFile}},
+	}, &got)
+
+	err := runInputCommand(t, cmd, "--content", "@   ")
+	if err == nil || !strings.Contains(err.Error(), "文件路径不能为空") {
+		t.Fatalf("expected empty-path rejection, got %v", err)
+	}
+}
+
+// A flag without Input keeps its literal value even when it looks like a
+// source reference — resolution is strictly opt-in per declaration.
+func TestCrossPlatformCoverageResolveInputFlagsNoInputSpecPassthrough(t *testing.T) {
+	var got map[string]any
+	cmd := newInputCommand([]FlagSpec{
+		{Name: "token", Usage: "T", Bind: "token"},
+	}, &got)
+
+	if err := runInputCommand(t, cmd, "--token", "@not-a-file"); err != nil {
+		t.Fatalf("execute: %v", err)
+	}
+	if got["token"] != "@not-a-file" {
+		t.Fatalf("toolArgs[token] = %q", got["token"])
+	}
+}
+
+// Registration defaults and env fallback are never input-resolved: @file only
+// applies to what the user literally typed on the command line.
+func TestCrossPlatformCoverageResolveInputFlagsDefaultNotResolved(t *testing.T) {
+	var got map[string]any
+	cmd := newInputCommand([]FlagSpec{
+		{Name: "content", Usage: "C", Bind: "content", Default: "@not-a-file", Input: []string{InputFile}},
+	}, &got)
+
+	if err := runInputCommand(t, cmd); err != nil {
+		t.Fatalf("execute: %v", err)
+	}
+	if got["content"] != "@not-a-file" {
+		t.Fatalf("toolArgs[content] = %q", got["content"])
+	}
+}
+
+// Trim flags judge usability on the trimmed value (rawValue), so a leading
+// whitespace before @ must still resolve instead of shipping as a literal.
+func TestCrossPlatformCoverageResolveInputFlagsTrimmedLeadingSpace(t *testing.T) {
+	var got map[string]any
+	cmd := newInputCommand([]FlagSpec{
+		{Name: "content", Usage: "C", Bind: "content", Trim: true, Input: []string{InputFile}},
+	}, &got)
+	path := writeInputFile(t, "trimmed payload")
+
+	if err := runInputCommand(t, cmd, "--content", " @"+path); err != nil {
+		t.Fatalf("execute: %v", err)
+	}
+	if got["content"] != "trimmed payload" {
+		t.Fatalf("toolArgs[content] = %q", got["content"])
+	}
+}
+
+func TestCrossPlatformCoverageValidateInputSpecsPanics(t *testing.T) {
+	cases := []struct {
+		name string
+		flag FlagSpec
+	}{
+		{"non-string kind", FlagSpec{Name: "n", Kind: KindInt, Input: []string{InputFile}}},
+		{"unknown source", FlagSpec{Name: "s", Input: []string{"url"}}},
+		{"duplicate source", FlagSpec{Name: "s", Input: []string{InputFile, InputFile}}},
+	}
+	for _, tc := range cases {
+		t.Run(tc.name, func(t *testing.T) {
+			defer func() {
+				if recover() == nil {
+					t.Fatal("expected panic")
+				}
+			}()
+			New(Spec{
+				Use:    "t",
+				Short:  "t",
+				Flags:  []FlagSpec{tc.flag},
+				Invoke: func(c *Ctx, toolArgs map[string]any) error { return nil },
+			})
+		})
+	}
+}

internal/corecmd/wait_phase_test.go

@@ -0,0 +1,902 @@
+// Copyright 2026 Alibaba Group
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+package corecmd
+
+import (
+	"bytes"
+	"context"
+	"errors"
+	"io"
+	"math"
+	"strconv"
+	"strings"
+	"testing"
+	"time"
+
+	"github.com/spf13/cobra"
+
+	"github.com/DingTalk-Real-AI/dingtalk-workspace-cli/internal/corecmd/contract"
+	"github.com/DingTalk-Real-AI/dingtalk-workspace-cli/internal/corecmd/contractfinal"
+	"github.com/DingTalk-Real-AI/dingtalk-workspace-cli/internal/output"
+	"github.com/DingTalk-Real-AI/dingtalk-workspace-cli/internal/wait"
+)
+
+func waitTestDecl() ContractDecl {
+	return ContractDecl{
+		Title:       "Wait Title",
+		Description: "Wait Desc",
+		Wait: &contract.WaitSpec{
+			Mode:               contract.WaitModePoll,
+			PollCommand:        "oa approval-instance get",
+			StatusQuery:        "result.status",
+			Terminal:           map[string]contract.ResultOutcome{"COMPLETED": contract.ResultOutcomeSuccess, "REJECTED": contract.ResultOutcomeFailure},
+			PendingValues:      []string{"NEW", "RUNNING"},
+			DefaultTimeoutSecs: 60,
+		},
+		Interface: &contract.InterfaceSpec{Mode: "local", Availability: "available"},
+		Selection: contract.SelectionSpec{
+			AgentSummary: "summary",
+			UseWhen:      []string{"when wait"},
+			AvoidWhen:    []string{"when nowait"},
+			Examples:     []string{"dws wait-sample --wait"},
+		},
+		Identity: contract.ToolIdentitySpec{ProductID: "sample", Name: "waitsample", CanonicalPath: "sample.waitsample", CLIPath: "wait-sample", PrimaryCLIPath: "wait-sample"},
+	}
+}
+
+func baseWaitSpec(decl ContractDecl, poll func(context.Context, *Ctx) (wait.PollDoc, error)) Spec {
+	return Spec{
+		Use:           "wait-sample",
+		OutputRollout: output.RolloutUnifiedActive,
+		Safety:        contract.SafetySpec{Effect: "read", Risk: "low", Confirmation: "not_required", Idempotency: "idempotent"},
+		Contract:      decl,
+		WaitPoll:      poll,
+		ResultInvoke: func(*Ctx, map[string]any) (output.CommandResult, error) {
+			return output.Pending(map[string]any{"id": "job-1"}, &output.OperationInfo{
+				ID:          "job-1",
+				State:       "NEW",
+				NextCommand: "dws wait-sample --id job-1",
+			}), nil
+		},
+	}
+}
+
+func TestWaitFlagsOnlyRegisteredWhenDeclared(t *testing.T) {
+	declared := New(baseWaitSpec(waitTestDecl(), func(context.Context, *Ctx) (wait.PollDoc, error) {
+		return wait.PollDoc{"result": map[string]any{"status": "COMPLETED"}}, nil
+	}))
+	if flag := declared.Flags().Lookup(waitFlagName); flag == nil {
+		t.Fatal("declared command missing --wait flag")
+	}
+	if flag := declared.Flags().Lookup(waitTimeoutFlagName); flag == nil {
+		t.Fatal("declared command missing --wait-timeout flag")
+	}
+
+	undeclared := New(Spec{
+		Use:    "nowait",
+		Safety: contract.SafetySpec{Effect: "read", Risk: "low", Confirmation: "not_required", Idempotency: "idempotent"},
+		Invoke: func(*Ctx, map[string]any) error { return nil },
+	})
+	if flag := undeclared.Flags().Lookup(waitFlagName); flag != nil {
+		t.Fatal("undeclared command registered --wait")
+	}
+	undeclared.SetArgs([]string{"--wait"})
+	if err := undeclared.Execute(); err == nil || !strings.Contains(err.Error(), "unknown flag") {
+		t.Fatalf("err=%v want unknown-flag", err)
+	}
+}
+
+func TestValidateWaitDeclPairsDeclarationWithImplementation(t *testing.T) {
+	decl := waitTestDecl()
+	spec := baseWaitSpec(decl, nil)
+	expectPanic(t, func() { New(spec) }, "WaitPoll")
+
+	spec.WaitPoll = func(context.Context, *Ctx) (wait.PollDoc, error) { return nil, nil }
+	expectPanic(t, func() {
+		New(Spec{
+			Use:      "hook-only",
+			Safety:   contract.SafetySpec{Effect: "read", Risk: "low", Confirmation: "not_required", Idempotency: "idempotent"},
+			Invoke:   func(*Ctx, map[string]any) error { return nil },
+			WaitPoll: func(context.Context, *Ctx) (wait.PollDoc, error) { return nil, nil },
+		})
+	}, "Contract.Wait")
+}
+
+func expectPanic(t *testing.T, fn func(), want string) {
+	t.Helper()
+	defer func() {
+		recovered := recover()
+		if recovered == nil {
+			t.Fatalf("expected panic containing %q", want)
+		}
+		if message, ok := recovered.(string); !ok || !strings.Contains(message, want) {
+			t.Fatalf("panic=%v want containing %q", recovered, want)
+		}
+	}()
+	fn()
+}
+
+func TestWaitTimeoutFlagDefaultsComeFromDeclaration(t *testing.T) {
+	stub := func(context.Context, *Ctx) (wait.PollDoc, error) {
+		return wait.PollDoc{"result": map[string]any{"status": "COMPLETED"}}, nil
+	}
+	declared := New(baseWaitSpec(waitTestDecl(), stub))
+	if value, err := declared.Flags().GetInt(waitTimeoutFlagName); err != nil || value != 60 {
+		t.Fatalf("declared default=%d/%v, want reviewed 60", value, err)
+	}
+	decl := waitTestDecl()
+	decl.Wait.DefaultTimeoutSecs = 0
+	fallback := New(baseWaitSpec(decl, stub))
+	if value, err := fallback.Flags().GetInt(waitTimeoutFlagName); err != nil || value != DefaultWaitTimeoutSecs {
+		t.Fatalf("fallback default=%d/%v, want framework %d", value, err, DefaultWaitTimeoutSecs)
+	}
+	// A non-positive explicit value falls back to the framework default.
+	fallback.SetArgs([]string{"--wait-timeout", "0", "--wait"})
+	if err := fallback.Flags().Set(waitTimeoutFlagName, "0"); err != nil {
+		t.Fatal(err)
+	}
+	if got := waitTimeoutSecs(fallback); got != DefaultWaitTimeoutSecs {
+		t.Fatalf("waitTimeoutSecs=%d, want %d", got, DefaultWaitTimeoutSecs)
+	}
+}
+
+func TestWaitTimeoutDurationRejectsOverflowingSeconds(t *testing.T) {
+	// math.MaxInt64 (9223372036854775807) is a legal pflag int on 64-bit
+	// platforms and overflows time.Duration(secs)*time.Second to a negative
+	// value, which would disable the wait deadline.
+	if _, err := waitTimeoutDuration(math.MaxInt64); err == nil || !strings.Contains(err.Error(), "超出可表示范围") {
+		t.Fatalf("err=%v, want overflow validation", err)
+	}
+	d, err := waitTimeoutDuration(maxWaitTimeoutSecs)
+	if err != nil {
+		t.Fatal(err)
+	}
+	if d <= 0 || d != time.Duration(maxWaitTimeoutSecs)*time.Second {
+		t.Fatalf("duration=%d, want the largest representable timeout", d)
+	}
+	// Non-positive second counts fall back to the framework default instead
+	// of disabling the deadline (waitTimeoutSecs already maps a zero/negative
+	// flag to the default; this keeps the conversion itself fail-safe).
+	if got, err := waitTimeoutDuration(0); err != nil || got != time.Duration(DefaultWaitTimeoutSecs)*time.Second {
+		t.Fatalf("duration/err=%d/%v, want framework default", got, err)
+	}
+	if got, err := waitTimeoutDuration(-5); err != nil || got != time.Duration(DefaultWaitTimeoutSecs)*time.Second {
+		t.Fatalf("duration/err=%d/%v, want framework default", got, err)
+	}
+}
+
+func TestResultInvokeWaitTimeoutOverflowIsValidationError(t *testing.T) {
+	if int64(math.MaxInt) <= maxWaitTimeoutSecs {
+		t.Skip("platform int cannot overflow time.Duration")
+	}
+	polled := false
+	cmd := New(baseWaitSpec(waitTestDecl(), func(context.Context, *Ctx) (wait.PollDoc, error) {
+		polled = true
+		return wait.PollDoc{"result": map[string]any{"status": "COMPLETED"}}, nil
+	}))
+	cmd.SetArgs([]string{"--wait", "--wait-timeout", strconv.Itoa(math.MaxInt)})
+	ctx, _ := output.WithResultStore(context.Background())
+	cmd.SetContext(ctx)
+	err := cmd.Execute()
+	if err == nil || !strings.Contains(err.Error(), "超出可表示范围") {
+		t.Fatalf("err=%v, want overflow validation", err)
+	}
+	if polled {
+		t.Fatal("overflowing --wait-timeout must not start the wait loop")
+	}
+}
+
+func TestResultInvokeWaitPollErrorFailsTheCommand(t *testing.T) {
+	boom := errors.New("rpc down")
+	cmd := New(baseWaitSpec(waitTestDecl(), func(context.Context, *Ctx) (wait.PollDoc, error) {
+		return nil, boom
+	}))
+	cmd.SetArgs([]string{"--wait"})
+	ctx, _ := output.WithResultStore(context.Background())
+	cmd.SetContext(ctx)
+	err := cmd.Execute()
+	if err == nil || !strings.Contains(err.Error(), "rpc down") {
+		t.Fatalf("err=%v, want poll error surfaced", err)
+	}
+}
+
+func TestResultInvokeWaitUnknownStatusFailsClosed(t *testing.T) {
+	cmd := New(baseWaitSpec(waitTestDecl(), func(context.Context, *Ctx) (wait.PollDoc, error) {
+		return wait.PollDoc{"result": map[string]any{"status": "Mystery"}}, nil
+	}))
+	cmd.SetArgs([]string{"--wait"})
+	ctx, _ := output.WithResultStore(context.Background())
+	cmd.SetContext(ctx)
+	err := cmd.Execute()
+	if err == nil || !wait.IsUnknownStatus(err) {
+		t.Fatalf("err=%v, want unknown-status", err)
+	}
+}
+
+func TestWaitCtxAccessorsExposeDeclaredCapability(t *testing.T) {
+	var gotWait bool
+	var gotTimeout int
+	cmd := New(baseWaitSpec(waitTestDecl(), func(_ context.Context, c *Ctx) (wait.PollDoc, error) {
+		gotWait = c.Wait()
+		gotTimeout = c.WaitTimeoutSecs()
+		return wait.PollDoc{"result": map[string]any{"status": "COMPLETED"}}, nil
+	}))
+	cmd.SetArgs([]string{"--wait", "--wait-timeout", "90"})
+	ctx, _ := output.WithResultStore(context.Background())
+	cmd.SetContext(ctx)
+	if err := cmd.Execute(); err != nil {
+		t.Fatal(err)
+	}
+	if !gotWait || gotTimeout != 90 {
+		t.Fatalf("ctx accessors=%v/%d", gotWait, gotTimeout)
+	}
+}
+
+func eventTestDecl(mode string) ContractDecl {
+	decl := waitTestDecl()
+	decl.Wait.Mode = mode
+	decl.Wait.EventKey = "bpms_instance_change"
+	decl.Wait.MatchField = "process_instance_id"
+	decl.Wait.ResourceQuery = "id"
+	return decl
+}
+
+type scriptedStream struct {
+	events []wait.PollDoc
+	err    error
+}
+
+func (s *scriptedStream) Recv(context.Context) (wait.PollDoc, error) {
+	if len(s.events) > 0 {
+		doc := s.events[0]
+		s.events = s.events[1:]
+		return doc, nil
+	}
+	if s.err != nil {
+		return nil, s.err
+	}
+	return nil, io.EOF
+}
+
+func TestValidateWaitDeclPairsModeWithHooks(t *testing.T) {
+	poll := func(context.Context, *Ctx) (wait.PollDoc, error) { return nil, nil }
+	events := func(context.Context, *Ctx) (wait.EventStream, error) { return nil, nil }
+	cases := []struct {
+		name       string
+		mode       string
+		waitPoll   bool
+		waitEvents bool
+		want       string
+	}{
+		{"event without WaitEvents", contract.WaitModeEvent, false, false, "WaitEvents"},
+		{"auto without WaitPoll", contract.WaitModeAuto, false, true, "WaitPoll"},
+		{"poll with WaitEvents", contract.WaitModePoll, true, true, "WaitEvents"},
+		{"event with WaitPoll", contract.WaitModeEvent, true, true, "WaitPoll"},
+	}
+	for _, tc := range cases {
+		t.Run(tc.name, func(t *testing.T) {
+			expectPanic(t, func() {
+				New(Spec{
+					Use:           "wait-sample",
+					OutputRollout: output.RolloutUnifiedActive,
+					Safety:        contract.SafetySpec{Effect: "read", Risk: "low", Confirmation: "not_required", Idempotency: "idempotent"},
+					Contract:      eventTestDecl(tc.mode),
+					ResultInvoke: func(*Ctx, map[string]any) (output.CommandResult, error) {
+						return output.Pending(map[string]any{}, nil), nil
+					},
+					WaitPoll:   hookOrNil(tc.waitPoll, poll),
+					WaitEvents: eventHookOrNil(tc.waitEvents, events),
+				})
+			}, tc.want)
+		})
+	}
+}
+
+func hookOrNil(set bool, hook func(context.Context, *Ctx) (wait.PollDoc, error)) func(context.Context, *Ctx) (wait.PollDoc, error) {
+	if !set {
+		return nil
+	}
+	return hook
+}
+
+func eventHookOrNil(set bool, hook func(context.Context, *Ctx) (wait.EventStream, error)) func(context.Context, *Ctx) (wait.EventStream, error) {
+	if !set {
+		return nil
+	}
+	return hook
+}
+
+func runWaitModeCommand(t *testing.T, decl ContractDecl, poll func(context.Context, *Ctx) (wait.PollDoc, error), events func(context.Context, *Ctx) (wait.EventStream, error), args ...string) (string, error) {
+	t.Helper()
+	cmd := New(Spec{
+		Use:           "wait-sample",
+		OutputRollout: output.RolloutUnifiedActive,
+		Safety:        contract.SafetySpec{Effect: "read", Risk: "low", Confirmation: "not_required", Idempotency: "idempotent"},
+		Contract:      decl,
+		ResultInvoke: func(*Ctx, map[string]any) (output.CommandResult, error) {
+			return output.Pending(map[string]any{"id": "job-1"}, &output.OperationInfo{
+				ID: "job-1", State: "NEW", NextCommand: "dws wait-sample --id job-1",
+			}), nil
+		},
+		WaitPoll:   poll,
+		WaitEvents: events,
+	})
+	cmd.SetArgs(append([]string{"--wait"}, args...))
+	ctx, _ := output.WithResultStore(context.Background())
+	cmd.SetContext(ctx)
+	var stdout bytes.Buffer
+	cmd.SetOut(&stdout)
+	cmd.PersistentPostRunE = func(executed *cobra.Command, _ []string) error {
+		_, _, err := output.EmitStoredResult(executed)
+		return err
+	}
+	err := cmd.Execute()
+	return stdout.String(), err
+}
+
+func TestEventModeClosesEnvelopeFromCorrelatedEvent(t *testing.T) {
+	stream := &scriptedStream{events: []wait.PollDoc{
+		{"process_instance_id": "other", "result": map[string]any{"status": "COMPLETED"}},
+		{"process_instance_id": "job-1", "result": map[string]any{"status": "REJECTED"}},
+	}}
+	stdout, err := runWaitModeCommand(t, eventTestDecl(contract.WaitModeEvent), nil, func(context.Context, *Ctx) (wait.EventStream, error) {
+		return stream, nil
+	})
+	if err != nil {
+		t.Fatal(err)
+	}
+	if !strings.Contains(stdout, `"outcome": "failure"`) || !strings.Contains(stdout, `"type": "wait"`) {
+		t.Fatalf("stdout=%s", stdout)
+	}
+}
+
+func TestEventModeSurfacesStreamEndAsError(t *testing.T) {
+	_, err := runWaitModeCommand(t, eventTestDecl(contract.WaitModeEvent), nil, func(context.Context, *Ctx) (wait.EventStream, error) {
+		return &scriptedStream{}, nil
+	})
+	if err == nil || !errors.Is(err, wait.ErrEventStreamEnded) {
+		t.Fatalf("err=%v, want stream-ended", err)
+	}
+}
+
+func TestEventModeRejectsUnresolvableResource(t *testing.T) {
+	decl := eventTestDecl(contract.WaitModeEvent)
+	decl.Wait.ResourceQuery = "missing"
+	_, err := runWaitModeCommand(t, decl, nil, func(context.Context, *Ctx) (wait.EventStream, error) {
+		return &scriptedStream{}, nil
+	})
+	if err == nil || !strings.Contains(err.Error(), "resource query") {
+		t.Fatalf("err=%v", err)
+	}
+}
+
+func TestAutoModeFallsBackToPollOnStreamEnd(t *testing.T) {
+	polled := false
+	stdout, err := runWaitModeCommand(t, eventTestDecl(contract.WaitModeAuto),
+		func(context.Context, *Ctx) (wait.PollDoc, error) {
+			polled = true
+			return wait.PollDoc{"result": map[string]any{"status": "COMPLETED"}}, nil
+		},
+		func(context.Context, *Ctx) (wait.EventStream, error) {
+			return &scriptedStream{}, nil // ends immediately
+		})
+	if err != nil {
+		t.Fatal(err)
+	}
+	if !polled {
+		t.Fatal("auto mode did not fall back to polling")
+	}
+	if !strings.Contains(stdout, `"outcome": "success"`) {
+		t.Fatalf("stdout=%s", stdout)
+	}
+}
+
+func TestAutoModeFallsBackToPollOnSubscriptionFailure(t *testing.T) {
+	polled := false
+	_, err := runWaitModeCommand(t, eventTestDecl(contract.WaitModeAuto),
+		func(context.Context, *Ctx) (wait.PollDoc, error) {
+			polled = true
+			return wait.PollDoc{"result": map[string]any{"status": "COMPLETED"}}, nil
+		},
+		func(context.Context, *Ctx) (wait.EventStream, error) {
+			return nil, errors.New("no subscriber credential")
+		})
+	if err != nil {
+		t.Fatal(err)
+	}
+	if !polled {
+		t.Fatal("auto mode did not fall back to polling on subscription failure")
+	}
+}
+
+func TestResultInvokeWaitClosesEnvelopeOutcome(t *testing.T) {
+	cmd := New(baseWaitSpec(waitTestDecl(), func(context.Context, *Ctx) (wait.PollDoc, error) {
+		return wait.PollDoc{"result": map[string]any{"status": "REJECTED"}}, nil
+	}))
+	cmd.SetArgs([]string{"--wait"})
+	ctx, store := output.WithResultStore(context.Background())
+	cmd.SetContext(ctx)
+	var stdout bytes.Buffer
+	cmd.SetOut(&stdout)
+	cmd.PersistentPostRunE = func(executed *cobra.Command, _ []string) error {
+		_, _, err := output.EmitStoredResult(executed)
+		return err
+	}
+	if err := cmd.Execute(); err != nil {
+		t.Fatal(err)
+	}
+	if code, emitted := output.StoredExitCode(store); !emitted || code != 8 {
+		t.Fatalf("stored code/emitted=%d/%v, want dedicated wait-terminal-failure code 8", code, emitted)
+	}
+	if !strings.Contains(stdout.String(), `"type": "wait"`) {
+		t.Fatalf("stdout=%s, want error.type wait", stdout.String())
+	}
+	if !strings.Contains(stdout.String(), `"outcome": "failure"`) {
+		t.Fatalf("stdout=%s", stdout.String())
+	}
+	// The final emitted envelope must carry the observed terminal status in
+	// meta.operation.state — the acceptance-phase state ("NEW") must not
+	// survive the close (P1 regression guard).
+	if !strings.Contains(stdout.String(), `"state": "REJECTED"`) {
+		t.Fatalf("stdout=%s, want operation.state synced to the terminal status", stdout.String())
+	}
+	if strings.Contains(stdout.String(), `"state": "NEW"`) {
+		t.Fatalf("stdout=%s, acceptance-phase operation.state leaked into the terminal envelope", stdout.String())
+	}
+	if strings.Contains(stdout.String(), `"timed_out": true`) {
+		t.Fatalf("stdout=%s, terminal close must not claim timed_out", stdout.String())
+	}
+}
+
+func TestResultInvokeWaitSuccessCloseSyncsOperationState(t *testing.T) {
+	cmd := New(baseWaitSpec(waitTestDecl(), func(context.Context, *Ctx) (wait.PollDoc, error) {
+		return wait.PollDoc{"result": map[string]any{"status": "COMPLETED"}}, nil
+	}))
+	cmd.SetArgs([]string{"--wait"})
+	ctx, store := output.WithResultStore(context.Background())
+	cmd.SetContext(ctx)
+	var stdout bytes.Buffer
+	cmd.SetOut(&stdout)
+	cmd.PersistentPostRunE = func(executed *cobra.Command, _ []string) error {
+		_, _, err := output.EmitStoredResult(executed)
+		return err
+	}
+	if err := cmd.Execute(); err != nil {
+		t.Fatal(err)
+	}
+	if code, emitted := output.StoredExitCode(store); !emitted || code != 0 {
+		t.Fatalf("stored code/emitted=%d/%v, want success exit 0", code, emitted)
+	}
+	if !strings.Contains(stdout.String(), `"outcome": "success"`) {
+		t.Fatalf("stdout=%s", stdout.String())
+	}
+	// Success close must publish the terminal status, never the stale
+	// acceptance-phase state (no outcome=success with state=processing/NEW).
+	if !strings.Contains(stdout.String(), `"state": "COMPLETED"`) {
+		t.Fatalf("stdout=%s, want operation.state synced to the terminal status", stdout.String())
+	}
+	if strings.Contains(stdout.String(), `"state": "NEW"`) {
+		t.Fatalf("stdout=%s, acceptance-phase operation.state leaked into the success envelope", stdout.String())
+	}
+	// Operation identity (id / next_command) survives the terminal close.
+	if !strings.Contains(stdout.String(), `"id": "job-1"`) || !strings.Contains(stdout.String(), `"next_command"`) {
+		t.Fatalf("stdout=%s, want operation id/next_command preserved", stdout.String())
+	}
+}
+
+func TestResultInvokeWaitTimeoutKeepsPending(t *testing.T) {
+	polls := 0
+	cmd := New(baseWaitSpec(waitTestDecl(), func(context.Context, *Ctx) (wait.PollDoc, error) {
+		polls++
+		return wait.PollDoc{"result": map[string]any{"status": "RUNNING"}}, nil
+	}))
+	cmd.SetArgs([]string{"--wait", "--wait-timeout", "1"})
+	ctx, store := output.WithResultStore(context.Background())
+	cmd.SetContext(ctx)
+	var stdout bytes.Buffer
+	cmd.SetOut(&stdout)
+	cmd.PersistentPostRunE = func(executed *cobra.Command, _ []string) error {
+		_, _, err := output.EmitStoredResult(executed)
+		return err
+	}
+	if err := cmd.Execute(); err != nil {
+		t.Fatalf("timeout wait must exit 0 (pending is not failure): %v", err)
+	}
+	if code, emitted := output.StoredExitCode(store); !emitted || code != 0 {
+		t.Fatalf("stored code/emitted=%d/%v", code, emitted)
+	}
+	if !strings.Contains(stdout.String(), `"outcome": "pending"`) {
+		t.Fatalf("stdout=%s", stdout.String())
+	}
+	if polls == 0 {
+		t.Fatal("wait phase never polled")
+	}
+}
+
+func TestWaitDeclRequiresResultInvokeDispatcher(t *testing.T) {
+	// A declared wait on the legacy Invoke path would observe a failure
+	// terminal while still exiting 0 — construction must reject it.
+	expectPanic(t, func() {
+		New(Spec{
+			Use:      "wait-sample",
+			Safety:   contract.SafetySpec{Effect: "read", Risk: "low", Confirmation: "not_required", Idempotency: "idempotent"},
+			Contract: waitTestDecl(),
+			WaitPoll: func(context.Context, *Ctx) (wait.PollDoc, error) { return nil, nil },
+			Invoke:   func(*Ctx, map[string]any) error { return nil },
+		})
+	}, "ResultInvoke")
+}
+
+func TestResultInvokeWithoutWaitFlagSkipsPhase(t *testing.T) {
+	polled := false
+	cmd := New(baseWaitSpec(waitTestDecl(), func(context.Context, *Ctx) (wait.PollDoc, error) {
+		polled = true
+		return wait.PollDoc{"result": map[string]any{"status": "COMPLETED"}}, nil
+	}))
+	cmd.SetArgs(nil)
+	ctx, _ := output.WithResultStore(context.Background())
+	cmd.SetContext(ctx)
+	if err := cmd.Execute(); err != nil {
+		t.Fatal(err)
+	}
+	if polled {
+		t.Fatal("wait phase ran without --wait")
+	}
+}
+
+func TestAttachContractPanicsOnInvalidWaitDeclaration(t *testing.T) {
+	decl := waitTestDecl()
+	decl.Wait.Mode = "event" // not implemented
+	defer func() {
+		recovered := recover()
+		if recovered == nil {
+			t.Fatal("expected panic on invalid Contract.Wait")
+		}
+		if message, ok := recovered.(string); !ok || !strings.Contains(message, "Contract.Wait") {
+			t.Fatalf("panic=%v", recovered)
+		}
+	}()
+	AttachContract(&cobra.Command{Use: "wait-sample"}, contract.SafetySpec{
+		Effect: "read", Risk: "low", Confirmation: "not_required", Idempotency: "idempotent",
+	}, decl, "", "")
+}
+
+func TestWaitDeclPaddedStatusValuesAreNormalized(t *testing.T) {
+	// A declaration whose status values carry surrounding whitespace must be
+	// canonicalized at construction so the runtime wait engine and the
+	// published Schema agree: the backend returns "COMPLETED" verbatim, and
+	// a padded terminal key would fail closed as an unknown status.
+	decl := waitTestDecl()
+	decl.Wait.Terminal = map[string]contract.ResultOutcome{
+		" COMPLETED ": contract.ResultOutcomeSuccess,
+		"\tREJECTED":  contract.ResultOutcomeFailure,
+	}
+	decl.Wait.PendingValues = []string{" NEW ", "RUNNING "}
+	cmd := New(baseWaitSpec(decl, func(context.Context, *Ctx) (wait.PollDoc, error) {
+		return wait.PollDoc{"result": map[string]any{"status": "COMPLETED"}}, nil
+	}))
+	cmd.SetArgs([]string{"--wait"})
+	ctx, store := output.WithResultStore(context.Background())
+	cmd.SetContext(ctx)
+	cmd.PersistentPostRunE = func(executed *cobra.Command, _ []string) error {
+		_, _, err := output.EmitStoredResult(executed)
+		return err
+	}
+	if err := cmd.Execute(); err != nil {
+		t.Fatalf("padded declaration must still reach the terminal status: %v", err)
+	}
+	if code, emitted := output.StoredExitCode(store); !emitted || code != 0 {
+		t.Fatalf("stored code/emitted=%d/%v, want success", code, emitted)
+	}
+	final, ok := contractfinal.RuntimeContractFinal(cmd)
+	if !ok || final.Wait == nil {
+		t.Fatal("registered ContractFinal lost the wait capability")
+	}
+	if _, ok := final.Wait.Terminal["COMPLETED"]; !ok {
+		t.Fatalf("registered terminal table not trimmed: %#v", final.Wait.Terminal)
+	}
+	for _, value := range final.Wait.PendingValues {
+		if strings.TrimSpace(value) != value {
+			t.Fatalf("registered pending value %q not trimmed", value)
+		}
+	}
+}
+
+func TestNewPanicsOnDuplicateOrConflictingWaitStatusesAfterTrim(t *testing.T) {
+	// Values that collapse onto one status after trimming are programming
+	// errors: silently merging them would pick one outcome for two authored
+	// declarations.
+	dupTerminal := waitTestDecl()
+	dupTerminal.Wait.Terminal = map[string]contract.ResultOutcome{
+		"COMPLETED":  contract.ResultOutcomeSuccess,
+		" COMPLETED": contract.ResultOutcomeFailure,
+	}
+	expectPanic(t, func() { New(baseWaitSpec(dupTerminal, nil)) }, "Contract.Wait")
+
+	conflict := waitTestDecl()
+	conflict.Wait.PendingValues = []string{" COMPLETED "}
+	expectPanic(t, func() { New(baseWaitSpec(conflict, nil)) }, "Contract.Wait")
+}
+
+func TestContractDeclEmptyTreatsWaitAsAuthored(t *testing.T) {
+	// Only Wait is authored: empty() must report non-empty through the Wait
+	// branch (before validateContractDecl then fails on the missing prose).
+	decl := ContractDecl{Wait: &contract.WaitSpec{
+		Mode:        contract.WaitModePoll,
+		PollCommand: "oa approval-instance get",
+		StatusQuery: "result.status",
+		Terminal:    map[string]contract.ResultOutcome{"COMPLETED": contract.ResultOutcomeSuccess},
+	}}
+	if decl.Empty() {
+		t.Fatal("Wait-only declaration must count as authored")
+	}
+	defer func() {
+		if recover() == nil {
+			t.Fatal("expected validateContractDecl to reject the missing prose")
+		}
+	}()
+	validateContractDecl(Spec{Use: "wait-only", Contract: decl})
+}
+
+func TestEventModeRejectsNonObjectResultData(t *testing.T) {
+	cmd := New(Spec{
+		Use:           "wait-sample",
+		OutputRollout: output.RolloutUnifiedActive,
+		Safety:        contract.SafetySpec{Effect: "read", Risk: "low", Confirmation: "not_required", Idempotency: "idempotent"},
+		Contract:      eventTestDecl(contract.WaitModeEvent),
+		ResultInvoke: func(*Ctx, map[string]any) (output.CommandResult, error) {
+			return output.Pending([]any{"not", "an", "object"}, &output.OperationInfo{
+				ID: "job-1", State: "NEW", NextCommand: "dws wait-sample",
+			}), nil
+		},
+		WaitEvents: func(context.Context, *Ctx) (wait.EventStream, error) {
+			return &scriptedStream{}, nil
+		},
+	})
+	cmd.SetArgs([]string{"--wait"})
+	ctx, _ := output.WithResultStore(context.Background())
+	cmd.SetContext(ctx)
+	err := cmd.Execute()
+	if err == nil || !strings.Contains(err.Error(), "not an object") {
+		t.Fatalf("err=%v, want non-object data rejection", err)
+	}
+}
+
+func TestEventModeSubscriptionFailureSurfacesInStrictMode(t *testing.T) {
+	decl := eventTestDecl(contract.WaitModeEvent)
+	_, err := runWaitModeCommand(t, decl, nil, func(context.Context, *Ctx) (wait.EventStream, error) {
+		return nil, errors.New("no subscriber credential")
+	})
+	if err == nil || !strings.Contains(err.Error(), "subscription failed") {
+		t.Fatalf("err=%v, want subscription failure surfaced", err)
+	}
+}
+
+func TestResultInvokeNonPendingSkipsWaitPhase(t *testing.T) {
+	partial, err := output.NewPartialData(2,
+		[]any{map[string]any{"id": "ok"}},
+		[]output.PartialFailedEntry{{ID: "bad", Error: &output.ErrorInfo{Type: "api", Message: "item failed"}}},
+		nil)
+	if err != nil {
+		t.Fatal(err)
+	}
+	cases := []struct {
+		name   string
+		result output.CommandResult
+		want   string
+	}{
+		{"failure", output.Failure(&output.ErrorInfo{Type: "api", Message: "business failed"}), "failure"},
+		{"success", output.Success(map[string]any{"id": "job-1"}), "success"},
+		{"partial", output.Partial(partial), "partial_failure"},
+	}
+	for _, tc := range cases {
+		t.Run(tc.name, func(t *testing.T) {
+			polled := false
+			subscribed := false
+			cmd := New(Spec{
+				Use:           "wait-sample",
+				OutputRollout: output.RolloutUnifiedActive,
+				Safety:        contract.SafetySpec{Effect: "read", Risk: "low", Confirmation: "not_required", Idempotency: "idempotent"},
+				Contract:      eventTestDecl(contract.WaitModeAuto),
+				ResultInvoke: func(*Ctx, map[string]any) (output.CommandResult, error) {
+					return tc.result, nil
+				},
+				WaitPoll: func(context.Context, *Ctx) (wait.PollDoc, error) {
+					polled = true
+					return wait.PollDoc{"result": map[string]any{"status": "COMPLETED"}}, nil
+				},
+				WaitEvents: func(context.Context, *Ctx) (wait.EventStream, error) {
+					subscribed = true
+					return &scriptedStream{}, nil
+				},
+			})
+			cmd.SetArgs([]string{"--wait"})
+			ctx, store := output.WithResultStore(context.Background())
+			cmd.SetContext(ctx)
+			var stdout bytes.Buffer
+			cmd.SetOut(&stdout)
+			cmd.PersistentPostRunE = func(executed *cobra.Command, _ []string) error {
+				_, _, err := output.EmitStoredResult(executed)
+				return err
+			}
+			if err := cmd.Execute(); err != nil {
+				t.Fatal(err)
+			}
+			if polled || subscribed {
+				t.Fatal("wait phase must not call WaitPoll/WaitEvents for a non-pending initial result")
+			}
+			if _, emitted := output.StoredExitCode(store); !emitted {
+				t.Fatal("initial result was not stored")
+			}
+			if !strings.Contains(stdout.String(), `"outcome": "`+tc.want+`"`) {
+				t.Fatalf("stdout=%s, want outcome %s preserved", stdout.String(), tc.want)
+			}
+			if strings.Contains(stdout.String(), `"type": "wait"`) {
+				t.Fatalf("stdout=%s, wait phase overwrote the original envelope", stdout.String())
+			}
+		})
+	}
+}
+
+func TestWaitTimeoutCancelsBlockingPoll(t *testing.T) {
+	started := make(chan struct{})
+	cmd := New(baseWaitSpec(waitTestDecl(), func(ctx context.Context, c *Ctx) (wait.PollDoc, error) {
+		close(started)
+		select {
+		case <-ctx.Done():
+			return nil, ctx.Err()
+		case <-c.Command().Context().Done():
+			return nil, c.Command().Context().Err()
+		}
+	}))
+	cmd.SetArgs([]string{"--wait", "--wait-timeout", "1"})
+	ctx, store := output.WithResultStore(context.Background())
+	cmd.SetContext(ctx)
+	var stdout bytes.Buffer
+	cmd.SetOut(&stdout)
+	cmd.PersistentPostRunE = func(executed *cobra.Command, _ []string) error {
+		_, _, err := output.EmitStoredResult(executed)
+		return err
+	}
+	done := make(chan error, 1)
+	go func() { done <- cmd.Execute() }()
+	select {
+	case <-started:
+	case <-time.After(2 * time.Second):
+		t.Fatal("blocking poll never started")
+	}
+	select {
+	case err := <-done:
+		if err != nil {
+			t.Fatalf("timeout wait must exit 0 (pending is not failure): %v", err)
+		}
+	case <-time.After(3 * time.Second):
+		t.Fatal("blocked poll was not cancelled by --wait-timeout")
+	}
+	if code, emitted := output.StoredExitCode(store); !emitted || code != 0 {
+		t.Fatalf("stored code/emitted=%d/%v", code, emitted)
+	}
+	if !strings.Contains(stdout.String(), `"outcome": "pending"`) {
+		t.Fatalf("stdout=%s", stdout.String())
+	}
+}
+
+func TestWaitTimeoutCancelsBlockingSubscribe(t *testing.T) {
+	started := make(chan struct{})
+	cmd := New(Spec{
+		Use:           "wait-sample",
+		OutputRollout: output.RolloutUnifiedActive,
+		Safety:        contract.SafetySpec{Effect: "read", Risk: "low", Confirmation: "not_required", Idempotency: "idempotent"},
+		Contract:      eventTestDecl(contract.WaitModeEvent),
+		ResultInvoke: func(*Ctx, map[string]any) (output.CommandResult, error) {
+			return output.Pending(map[string]any{"id": "job-1"}, &output.OperationInfo{
+				ID: "job-1", State: "NEW", NextCommand: "dws wait-sample --id job-1",
+			}), nil
+		},
+		WaitEvents: func(ctx context.Context, c *Ctx) (wait.EventStream, error) {
+			close(started)
+			// Leaf subscribe may wait on either the hook ctx or the cobra
+			// command context; both must carry the wait-timeout deadline.
+			select {
+			case <-ctx.Done():
+				return nil, ctx.Err()
+			case <-c.Command().Context().Done():
+				return nil, c.Command().Context().Err()
+			}
+		},
+	})
+	cmd.SetArgs([]string{"--wait", "--wait-timeout", "1"})
+	ctx, store := output.WithResultStore(context.Background())
+	cmd.SetContext(ctx)
+	var stdout bytes.Buffer
+	cmd.SetOut(&stdout)
+	cmd.PersistentPostRunE = func(executed *cobra.Command, _ []string) error {
+		_, _, err := output.EmitStoredResult(executed)
+		return err
+	}
+	done := make(chan error, 1)
+	go func() { done <- cmd.Execute() }()
+	select {
+	case <-started:
+	case <-time.After(2 * time.Second):
+		t.Fatal("blocking subscribe never started")
+	}
+	select {
+	case err := <-done:
+		if err != nil {
+			t.Fatalf("timeout wait must exit 0 (pending is not failure): %v", err)
+		}
+	case <-time.After(3 * time.Second):
+		t.Fatal("blocked subscribe was not cancelled by --wait-timeout")
+	}
+	if code, emitted := output.StoredExitCode(store); !emitted || code != 0 {
+		t.Fatalf("stored code/emitted=%d/%v", code, emitted)
+	}
+	if !strings.Contains(stdout.String(), `"outcome": "pending"`) {
+		t.Fatalf("stdout=%s", stdout.String())
+	}
+}
+
+func TestWaitTimeoutCancelsBlockingPollAfterAutoFallback(t *testing.T) {
+	started := make(chan struct{})
+	cmd := New(Spec{
+		Use:           "wait-sample",
+		OutputRollout: output.RolloutUnifiedActive,
+		Safety:        contract.SafetySpec{Effect: "read", Risk: "low", Confirmation: "not_required", Idempotency: "idempotent"},
+		Contract:      eventTestDecl(contract.WaitModeAuto),
+		ResultInvoke: func(*Ctx, map[string]any) (output.CommandResult, error) {
+			return output.Pending(map[string]any{"id": "job-1"}, &output.OperationInfo{
+				ID: "job-1", State: "NEW", NextCommand: "dws wait-sample --id job-1",
+			}), nil
+		},
+		WaitEvents: func(context.Context, *Ctx) (wait.EventStream, error) {
+			return &scriptedStream{}, nil // ends immediately → poll fallback
+		},
+		WaitPoll: func(ctx context.Context, _ *Ctx) (wait.PollDoc, error) {
+			close(started)
+			<-ctx.Done()
+			return nil, ctx.Err()
+		},
+	})
+	cmd.SetArgs([]string{"--wait", "--wait-timeout", "1"})
+	ctx, store := output.WithResultStore(context.Background())
+	cmd.SetContext(ctx)
+	var stdout bytes.Buffer
+	cmd.SetOut(&stdout)
+	cmd.PersistentPostRunE = func(executed *cobra.Command, _ []string) error {
+		_, _, err := output.EmitStoredResult(executed)
+		return err
+	}
+	done := make(chan error, 1)
+	go func() { done <- cmd.Execute() }()
+	select {
+	case <-started:
+	case <-time.After(2 * time.Second):
+		t.Fatal("auto-fallback blocking poll never started")
+	}
+	select {
+	case err := <-done:
+		if err != nil {
+			t.Fatalf("timeout wait must exit 0 (pending is not failure): %v", err)
+		}
+	case <-time.After(3 * time.Second):
+		t.Fatal("blocked auto-fallback poll was not cancelled by --wait-timeout")
+	}
+	if code, emitted := output.StoredExitCode(store); !emitted || code != 0 {
+		t.Fatalf("stored code/emitted=%d/%v", code, emitted)
+	}
+	if !strings.Contains(stdout.String(), `"outcome": "pending"`) {
+		t.Fatalf("stdout=%s", stdout.String())
+	}
+}

internal/errors/errors.go

@@ -58,6 +58,8 @@ const (
 //	5  internal       （CategoryInternal 与兜底：非结构化错误、panic 收敛均归 5）
 //	6  discovery      （CategoryDiscovery）
 //	7  partial_failure（部分成功专用码，见 ExitCodePartial）
+//	8  wait           （--wait 观察到失败终态的专用码，见 internal/output
+//	                   的 exitCodeWait；不设 Category，仅经统一信封产出）
 //
 // ExitCodePartial is the partial-result exit code shared with internal/output.
 // It is not returned for CategoryPartial errors because they lack the typed

internal/event/bus/coverage_edges_test.go

@@ -361,13 +361,18 @@ func TestCrossPlatformCoverageDaemonMethodEdges(t *testing.T) {
 		}
 		return nil, net.ErrClosed
 	}}
-	d := &daemon{listener: l, log: logger, hub: NewHub(1), idleStop: make(chan struct{})}
+	d := &daemon{listener: l, log: logger, hub: NewHub(1), idleStop: make(chan struct{}), stopReq: make(chan string, 1)}
 	d.acceptLoop(context.Background())
 	d.shuttingDown.Store(true)
 	d.acceptLoop(context.Background())
 	d.triggerShutdown("test")
-	if !l.closed {
-		t.Fatal("trigger did not close listener")
+	select {
+	case reason := <-d.stopReq:
+		if reason != "test" {
+			t.Fatalf("stop reason = %q", reason)
+		}
+	default:
+		t.Fatal("trigger did not notify lifecycle loop")
 	}
 
 	d = &daemon{listener: &scriptedListener{accept: func() (net.Conn, error) { return nil, net.ErrClosed }}, log: logger, hub: NewHub(1), idleStop: make(chan struct{})}

internal/event/bus/daemon.go

@@ -186,6 +186,7 @@ func Run(ctx context.Context, cfg Config) error {
 		dedup:    dd,
 		started:  time.Now().UTC(),
 		idleStop: make(chan struct{}),
+		stopReq:  make(chan string, 1),
 	}
 
 	// 4. Signal ready BEFORE accepting consumers (avoids a slow-fork
@@ -263,6 +264,9 @@ func Run(ctx context.Context, cfg Config) error {
 		}
 	case <-d.idleStop:
 		log.Info("bus: idle timeout reached, shutting down")
+	case reason := <-d.stopReq:
+		shutdownReason = reason
+		log.Info("bus: shutdown requested via IPC", "reason", reason)
 	}
 
 	// 7. Graceful shutdown — cancel runCtx first so all background
@@ -292,6 +296,7 @@ type daemon struct {
 	shutdownMu   sync.Mutex
 	shuttingDown atomic.Bool
 	idleStop     chan struct{}
+	stopReq      chan string
 
 	credentialHandoffMu sync.Mutex
 	terminalMu          sync.RWMutex
@@ -749,15 +754,17 @@ func (d *daemon) idleWatch(ctx context.Context) {
 	}
 }
 
-// triggerShutdown is called from RPC handlers that want to end the bus.
-// It works by closing the listener (which unblocks Run's select via the
-// source error path, indirectly). For v1 a full ctx-cancellation hook is
-// out of scope; busctl/stop also sends SIGTERM which is the authoritative
-// shutdown path.
+// triggerShutdown is called from RPC handlers that want to end the bus. The
+// buffered request wakes Run's lifecycle select, which then cancels the cloud
+// source and performs the same graceful shutdown used for signals and idle
+// expiry. It is intentionally non-blocking so repeated stop requests cannot
+// strand connection handlers.
 func (d *daemon) triggerShutdown(reason string) {
 	d.log.Info("bus: shutdown triggered via IPC", "reason", reason)
-	_ = d.listener.Close() // unblocks Accept(), but doesn't kill Source
-	// Best-effort: a future version wires a context.CancelFunc here.
+	select {
+	case d.stopReq <- reason:
+	default:
+	}
 }
 
 // shutdown performs the graceful tear-down sequence:

internal/event/bus/daemon_test.go

@@ -141,6 +141,67 @@ func TestDaemon_RunStartsAndShutsDownCleanly(t *testing.T) {
 	}
 }
 
+func TestCrossPlatformCoverageDaemonIPCStopShutsDownCleanly(t *testing.T) {
+	workDir := shortTempDir(t)
+	identityHash := dwsevent.IdentityHash(workDir)
+	endpoint := dwsevent.IPCEndpoint(workDir, "open", dwsevent.SourceKindAppStream, identityHash)
+	runDone := make(chan error, 1)
+	go func() {
+		runDone <- Run(context.Background(), Config{
+			WorkDir:      workDir,
+			IPCEndpoint:  endpoint,
+			ClientID:     "ding_ipc_stop_test",
+			Edition:      "open",
+			SourceKind:   dwsevent.SourceKindAppStream,
+			IdentityHash: identityHash,
+			Source:       &fakeSource{},
+		})
+	}()
+
+	var conn net.Conn
+	deadline := time.Now().Add(3 * time.Second)
+	for time.Now().Before(deadline) {
+		var err error
+		conn, err = transport.Dial(endpoint)
+		if err == nil {
+			break
+		}
+		time.Sleep(10 * time.Millisecond)
+	}
+	if conn == nil {
+		t.Fatal("bus IPC endpoint did not become ready")
+	}
+	w := transport.NewWriter(conn)
+	r := transport.NewReader(conn)
+	if err := w.WriteJSON(transport.Hello{
+		Type:        transport.FrameTypeHello,
+		ConsumerPID: os.Getpid(),
+		Role:        transport.HelloRoleStop,
+	}); err != nil {
+		t.Fatalf("write stop hello: %v", err)
+	}
+	var bye transport.Bye
+	if err := r.ReadJSON(&bye); err != nil {
+		t.Fatalf("read stop response: %v", err)
+	}
+	_ = conn.Close()
+	if bye.Type != transport.FrameTypeBye || bye.Reason != "stop_request" {
+		t.Fatalf("stop response = %#v", bye)
+	}
+
+	select {
+	case err := <-runDone:
+		if err != nil {
+			t.Fatalf("Run returned after IPC stop: %v", err)
+		}
+	case <-time.After(3 * time.Second):
+		t.Fatal("Run did not return after IPC stop")
+	}
+	if pid := ReadHolderPID(filepath.Join(workDir, LockFileName)); pid != 0 {
+		t.Fatalf("bus lock retained pid %d after IPC stop", pid)
+	}
+}
+
 func TestDaemon_ConsumerReceivesEvents(t *testing.T) {
 	skipOnWindows(t, "uses Unix socket dial")
 	workDir := shortTempDir(t)

internal/event/bus/lockfile.go

@@ -114,6 +114,47 @@ func ReadHolderPID(path string) int {
 	return parsePID(b)
 }
 
+// ValidateHolderOwnership reports whether pid still owns the exclusive lock at
+// path. If the lock can be acquired, no daemon owns it: the PID payload is
+// stale and is cleared while the probe holds the lock. The PID is read again
+// after a busy result so a daemon replacement cannot be mistaken for the
+// original owner.
+//
+// Callers use this immediately before sending a process-level termination
+// fallback. A live PID by itself is not proof of ownership because operating
+// systems may reuse PIDs after an unclean daemon exit.
+func ValidateHolderOwnership(path string, pid int) (bool, error) {
+	if pid <= 0 || ReadHolderPID(path) != pid {
+		return false, nil
+	}
+
+	probe, err := busTryAcquire(path)
+	if err != nil {
+		if errors.Is(err, ErrBusy) {
+			return ReadHolderPID(path) == pid, nil
+		}
+		return false, fmt.Errorf("bus: verify lock owner: %w", err)
+	}
+	defer probe.Close()
+
+	// We own the lock, so the recorded PID cannot own it. Clear only the PID
+	// we validated above; a changed payload is left untouched defensively.
+	f := probe.File()
+	if _, err := busSeek(f, 0, io.SeekStart); err != nil {
+		return false, fmt.Errorf("bus: seek stale lock: %w", err)
+	}
+	current, err := busReadAll(f)
+	if err != nil {
+		return false, fmt.Errorf("bus: read stale lock: %w", err)
+	}
+	if parsePID(current) == pid {
+		if err := truncateAndWritePID(f, 0); err != nil {
+			return false, fmt.Errorf("bus: clear stale PID: %w", err)
+		}
+	}
+	return false, nil
+}
+
 // Close releases the flock and best-effort blanks the PID body so a stale
 // reader (e.g. `event status` racing our shutdown) does not see our
 // long-dead PID and try to signal it. The lock file itself is NOT removed

internal/event/bus/lockfile_test.go

@@ -16,9 +16,13 @@ package bus
 import (
 	"errors"
 	"fmt"
+	"io"
 	"os"
 	"path/filepath"
 	"testing"
+
+	eventlock "github.com/DingTalk-Real-AI/dingtalk-workspace-cli/internal/event/lock"
+	"github.com/DingTalk-Real-AI/dingtalk-workspace-cli/internal/testseam"
 )
 
 func TestAcquire_WritesOurPID(t *testing.T) {
@@ -112,6 +116,109 @@ func TestReadHolderPID_MalformedReturnsZero(t *testing.T) {
 	}
 }
 
+func TestCrossPlatformCoverageValidateHolderOwnershipHeldLock(t *testing.T) {
+	path := filepath.Join(t.TempDir(), LockFileName)
+	held, err := busTryAcquire(path)
+	if err != nil {
+		t.Fatalf("hold lock: %v", err)
+	}
+	defer held.Close()
+	if err := truncateAndWritePID(held.File(), os.Getpid()); err != nil {
+		t.Fatalf("write holder PID: %v", err)
+	}
+
+	owned, err := ValidateHolderOwnership(path, os.Getpid())
+	if err != nil {
+		t.Fatalf("ValidateHolderOwnership: %v", err)
+	}
+	if !owned {
+		t.Fatal("held lock was not attributed to its recorded PID")
+	}
+}
+
+func TestCrossPlatformCoverageValidateHolderOwnershipClearsReusedPID(t *testing.T) {
+	path := filepath.Join(t.TempDir(), LockFileName)
+	// The current PID is alive but deliberately does not hold this lock. This
+	// models a stale daemon PID that the OS has reassigned to another process.
+	if err := os.WriteFile(path, []byte(fmt.Sprintf("%d\n", os.Getpid())), 0o600); err != nil {
+		t.Fatalf("write stale PID: %v", err)
+	}
+
+	owned, err := ValidateHolderOwnership(path, os.Getpid())
+	if err != nil {
+		t.Fatalf("ValidateHolderOwnership: %v", err)
+	}
+	if owned {
+		t.Fatal("live reused PID without the bus lock was accepted as owner")
+	}
+	if got := ReadHolderPID(path); got != 0 {
+		t.Fatalf("stale PID after validation = %d, want cleared", got)
+	}
+}
+
+func TestCrossPlatformCoverageValidateHolderOwnershipRejectsMismatch(t *testing.T) {
+	path := filepath.Join(t.TempDir(), LockFileName)
+	if err := os.WriteFile(path, []byte("123\n"), 0o600); err != nil {
+		t.Fatal(err)
+	}
+	owned, err := ValidateHolderOwnership(path, 456)
+	if err != nil || owned {
+		t.Fatalf("ValidateHolderOwnership mismatch = %v, %v", owned, err)
+	}
+}
+
+func TestCrossPlatformCoverageValidateHolderOwnershipErrors(t *testing.T) {
+	errInjected := errors.New("injected ownership validation failure")
+	writeLivePID := func(t *testing.T) string {
+		t.Helper()
+		path := filepath.Join(t.TempDir(), LockFileName)
+		if err := os.WriteFile(path, []byte(fmt.Sprintf("%d\n", os.Getpid())), 0o600); err != nil {
+			t.Fatal(err)
+		}
+		return path
+	}
+
+	t.Run("probe", func(t *testing.T) {
+		path := writeLivePID(t)
+		testseam.Swap(t, &busTryAcquire, func(string) (*eventlock.File, error) {
+			return nil, errInjected
+		})
+		if _, err := ValidateHolderOwnership(path, os.Getpid()); !errors.Is(err, errInjected) {
+			t.Fatalf("probe error = %v", err)
+		}
+	})
+
+	t.Run("seek", func(t *testing.T) {
+		path := writeLivePID(t)
+		testseam.Swap(t, &busSeek, func(*os.File, int64, int) (int64, error) {
+			return 0, errInjected
+		})
+		if _, err := ValidateHolderOwnership(path, os.Getpid()); !errors.Is(err, errInjected) {
+			t.Fatalf("seek error = %v", err)
+		}
+	})
+
+	t.Run("read", func(t *testing.T) {
+		path := writeLivePID(t)
+		testseam.Swap(t, &busReadAll, func(io.Reader) ([]byte, error) {
+			return nil, errInjected
+		})
+		if _, err := ValidateHolderOwnership(path, os.Getpid()); !errors.Is(err, errInjected) {
+			t.Fatalf("read error = %v", err)
+		}
+	})
+
+	t.Run("clear", func(t *testing.T) {
+		path := writeLivePID(t)
+		testseam.Swap(t, &busTruncate, func(*os.File, int64) error {
+			return errInjected
+		})
+		if _, err := ValidateHolderOwnership(path, os.Getpid()); !errors.Is(err, errInjected) {
+			t.Fatalf("clear error = %v", err)
+		}
+	})
+}
+
 func TestAcquire_AfterReleaseReclaimable(t *testing.T) {
 	path := filepath.Join(t.TempDir(), LockFileName)
 	l1, err := Acquire(path)

internal/event/busctl/coverage_edges_test.go

@@ -284,11 +284,13 @@ func TestCrossPlatformCoverageQueryStatusAndEntryEdges(t *testing.T) {
 
 func TestCrossPlatformCoverageStopInjectedEdges(t *testing.T) {
 	origRead := stopReadHolderPID
+	origValidateOwner := stopValidateHolderOwner
 	origAlive := stopAlive
 	origFind := stopFindProcess
 	origSignal := stopSignalProcess
 	t.Cleanup(func() {
 		stopReadHolderPID = origRead
+		stopValidateHolderOwner = origValidateOwner
 		stopAlive = origAlive
 		stopFindProcess = origFind
 		stopSignalProcess = origSignal
@@ -307,6 +309,7 @@ func TestCrossPlatformCoverageStopInjectedEdges(t *testing.T) {
 		t.Fatal(err)
 	}
 	stopFindProcess = func(int) (*os.Process, error) { return proc, nil }
+	stopValidateHolderOwner = func(string, int) (bool, error) { return true, nil }
 	stopSignalProcess = func(*os.Process, os.Signal) error { return os.ErrProcessDone }
 	if err := Stop(StopConfig{WorkDir: "x"}); err != nil {
 		t.Fatalf("done signal = %v", err)

internal/event/busctl/discover.go

@@ -112,8 +112,9 @@ func Discover(cfg DiscoverConfig) (net.Conn, error) {
 		return nil, fmt.Errorf("busctl: mkdir workdir: %w", err)
 	}
 	_, spawnErr := cfg.Spawn(SpawnConfig{
-		ClientID:  cfg.ClientID,
-		ExtraArgs: cfg.SpawnExtraArgs,
+		ClientID:    cfg.ClientID,
+		IPCEndpoint: cfg.IPCEndpoint,
+		ExtraArgs:   cfg.SpawnExtraArgs,
 	})
 	if spawnErr != nil && !errors.Is(spawnErr, ErrSpawnFailed) {
 		// Hard error (couldn't even exec the child). Stop here — no bus

internal/event/busctl/discover_test.go

@@ -122,7 +122,10 @@ func TestDiscover_NoBus_SpawnSucceeds(t *testing.T) {
 			closer()
 		}
 	})
-	fakeSpawn := func(SpawnConfig) (int, error) {
+	fakeSpawn := func(cfg SpawnConfig) (int, error) {
+		if cfg.IPCEndpoint != sock {
+			t.Fatalf("spawn IPC endpoint = %q, want %q", cfg.IPCEndpoint, sock)
+		}
 		closer = startStubBus(t, sock)
 		return 12345, nil
 	}

internal/event/busctl/spawn.go

@@ -43,18 +43,14 @@ var (
 	spawnPipe       = os.Pipe
 )
 
-// ErrSpawnFailed is returned when the child reports startup failure via
-// the ready pipe ('E' byte). The child's exit error / log file holds the
-// actual cause; this sentinel just lets the caller distinguish "ready
-// pipe said no" from "ready pipe timed out / closed early".
-var ErrSpawnFailed = errors.New("busctl: bus child reported startup failure on ready pipe")
+// ErrSpawnFailed is returned when the child reports startup failure through
+// the Unix ready pipe or exits before binding the Windows named pipe.
+var ErrSpawnFailed = errors.New("busctl: bus child reported startup failure")
 
 // ErrSpawnTimeout is returned when ReadyTimeout elapses without any signal.
 var ErrSpawnTimeout = errors.New("busctl: bus child did not signal readiness within deadline")
 
-// SpawnConfig describes one spawn attempt. ClientID is the only field
-// inspected by the child; the rest govern process attributes the parent
-// applies before exec.
+// SpawnConfig describes one spawn attempt.
 type SpawnConfig struct {
 	// ExecPath is the dws binary to exec. Default os.Executable().
 	ExecPath string
@@ -62,18 +58,23 @@ type SpawnConfig struct {
 	// ClientID is passed as `--client-id` to `dws event _bus`. Required.
 	ClientID string
 
+	// IPCEndpoint is the bus endpoint the child binds. Windows uses the
+	// existing named pipe as its readiness handshake because os/exec does not
+	// support ExtraFiles there. Unix keeps the inherited ready-pipe protocol.
+	IPCEndpoint string
+
 	// ExtraArgs are appended after `--client-id`. Empty for normal use; tests
 	// pass `--extra-flag-for-test` etc.
 	ExtraArgs []string
 
-	// Env to pass to the child. Defaults to os.Environ(). The ReadyFDEnv
-	// entry is appended automatically.
+	// Env to pass to the child. Defaults to os.Environ(). Unix appends the
+	// ReadyFDEnv entry automatically; Windows removes any inherited copy.
 	Env []string
 }
 
-// Spawn forks a detached `dws event _bus --client-id <id>` child process and
-// waits for it to signal readiness via the ready pipe. Returns the child's
-// PID on success — the caller can then dial the bus IPC endpoint.
+// spawnWithReadyPipe is the Unix implementation behind Spawn. It forks a
+// detached `dws event _bus --client-id <id>` child and waits for the inherited
+// ready pipe. Windows has a named-pipe implementation in spawn_windows.go.
 //
 // stdio detach (plan invariant #7):
 //   - cmd.Stdout / cmd.Stderr set to nil so the child's own writes don't
@@ -89,7 +90,7 @@ type SpawnConfig struct {
 // Parent (this function):
 //   - Holds the read end open until either 1 byte is read or ReadyTimeout
 //   - Returns ErrSpawnFailed for 'E', ErrSpawnTimeout otherwise
-func Spawn(cfg SpawnConfig) (pid int, err error) {
+func spawnWithReadyPipe(cfg SpawnConfig) (pid int, err error) {
 	if cfg.ClientID == "" {
 		return 0, errors.New("busctl: SpawnConfig.ClientID is required")
 	}

internal/event/busctl/spawn_test.go

@@ -22,6 +22,8 @@ import (
 	"strings"
 	"testing"
 	"time"
+
+	"github.com/DingTalk-Real-AI/dingtalk-workspace-cli/internal/event/transport"
 )
 
 // Test child mode: when DWS_BUSCTL_TEST_CHILD is set, this test binary
@@ -35,7 +37,8 @@ import (
 // env-marker pattern is what Go's own os/exec tests use and stays
 // confined to this file.
 const (
-	childEnvMarker = "DWS_BUSCTL_TEST_CHILD"
+	childEnvMarker   = "DWS_BUSCTL_TEST_CHILD"
+	childEndpointEnv = "DWS_BUSCTL_TEST_ENDPOINT"
 	// values:
 	//   "ready"        — write 'R' then sleep 30s (parent should see ready)
 	//   "fail"         — write 'E' then exit (parent should see ErrSpawnFailed)
@@ -70,6 +73,26 @@ func TestMain(m *testing.M) {
 		writeReady('R')
 		time.Sleep(30 * time.Second)
 		os.Exit(0)
+	case "windows-ready":
+		listener, err := transport.Listen(os.Getenv(childEndpointEnv))
+		if err != nil {
+			os.Exit(3)
+		}
+		defer listener.Close()
+		conn, err := listener.Accept()
+		if err != nil {
+			os.Exit(4)
+		}
+		_ = conn.Close()
+		time.Sleep(30 * time.Second)
+		os.Exit(0)
+	case "windows-fail":
+		os.Exit(5)
+	case "windows-exit":
+		os.Exit(0)
+	case "windows-stall":
+		time.Sleep(30 * time.Second)
+		os.Exit(0)
 	}
 	os.Exit(m.Run())
 }

internal/event/busctl/spawn_unix.go

@@ -20,6 +20,13 @@ import (
 	"syscall"
 )
 
+// Spawn starts the detached bus and waits for its inherited ready pipe.
+// ExtraFiles is intentionally confined to Unix: Go does not support it on
+// Windows.
+func Spawn(cfg SpawnConfig) (pid int, err error) {
+	return spawnWithReadyPipe(cfg)
+}
+
 // applyDetach configures the child to live past parent death and not share
 // the parent's controlling terminal. Setsid puts the child in a new
 // session, so SIGHUP on the parent's controlling tty (e.g. SSH disconnect)

internal/event/busctl/spawn_windows.go

@@ -16,14 +16,101 @@
 package busctl
 
 import (
+	"fmt"
+	"os"
 	"os/exec"
+	"strings"
 	"syscall"
+	"time"
+
+	"github.com/DingTalk-Real-AI/dingtalk-workspace-cli/internal/event/transport"
 )
 
 // CREATE_NEW_PROCESS_GROUP (0x00000200) prevents the child from receiving
 // the parent's Ctrl+C signal, similar in spirit to Setsid on Unix.
 const createNewProcessGroup = 0x00000200
 
+var (
+	spawnWindowsDial         = transport.Dial
+	spawnWindowsPollInterval = 25 * time.Millisecond
+)
+
+// Spawn starts a detached Windows bus without cmd.ExtraFiles (unsupported by
+// Go on Windows). Readiness is confirmed by dialing the bus's existing named
+// pipe. The child is reaped in the background after readiness, while an early
+// process exit is surfaced as ErrSpawnFailed.
+func Spawn(cfg SpawnConfig) (pid int, err error) {
+	if strings.TrimSpace(cfg.ClientID) == "" {
+		return 0, fmt.Errorf("busctl: SpawnConfig.ClientID is required")
+	}
+	if strings.TrimSpace(cfg.IPCEndpoint) == "" {
+		return 0, fmt.Errorf("busctl: SpawnConfig.IPCEndpoint is required on Windows")
+	}
+	if cfg.ExecPath == "" {
+		execPath, err := spawnExecutable()
+		if err != nil {
+			return 0, fmt.Errorf("busctl: locate executable: %w", err)
+		}
+		cfg.ExecPath = execPath
+	}
+	if cfg.Env == nil {
+		cfg.Env = os.Environ()
+	}
+
+	args := append([]string{"event", "_bus", "--client-id", cfg.ClientID}, cfg.ExtraArgs...)
+	cmd := exec.Command(cfg.ExecPath, args...)
+	cmd.Env = withoutReadyFDEnv(cfg.Env)
+	cmd.Stdin = nil
+	cmd.Stdout = nil
+	cmd.Stderr = nil
+	applyDetach(cmd)
+	if err := cmd.Start(); err != nil {
+		return 0, fmt.Errorf("busctl: start %s: %w", cfg.ExecPath, err)
+	}
+	pid = cmd.Process.Pid
+
+	waitDone := make(chan error, 1)
+	go func() { waitDone <- cmd.Wait() }()
+	timer := time.NewTimer(ReadyTimeout)
+	defer timer.Stop()
+	ticker := time.NewTicker(spawnWindowsPollInterval)
+	defer ticker.Stop()
+
+	for {
+		if conn, dialErr := spawnWindowsDial(cfg.IPCEndpoint); dialErr == nil {
+			_ = conn.Close()
+			return pid, nil
+		}
+		select {
+		case waitErr := <-waitDone:
+			if waitErr == nil {
+				return pid, fmt.Errorf("%w: child exited before binding named pipe", ErrSpawnFailed)
+			}
+			return pid, fmt.Errorf("%w: %v", ErrSpawnFailed, waitErr)
+		case <-timer.C:
+			_ = cmd.Process.Kill()
+			select {
+			case <-waitDone:
+			case <-time.After(time.Second):
+			}
+			return pid, ErrSpawnTimeout
+		case <-ticker.C:
+		}
+	}
+}
+
+func withoutReadyFDEnv(env []string) []string {
+	prefix := strings.ToUpper(ReadyFDEnv) + "="
+	out := make([]string, 0, len(env))
+	for _, entry := range env {
+		if strings.HasPrefix(strings.ToUpper(entry), prefix) {
+			continue
+		}
+		out = append(out, entry)
+	}
+	return out
+}
+
 func applyDetach(cmd *exec.Cmd) {
 	cmd.SysProcAttr = &syscall.SysProcAttr{
 		CreationFlags: createNewProcessGroup,