Merge pull request #1072 from DingTalk-Real-AI/codex/changelog-v1.0.59-stable

chore: prepare v1.0.59 changelog

.changes/released/1.0.59-beta.5/chat-personal-emotion.md

@@ -0,0 +1,5 @@
+---
+category: Added
+---
+
+- **Chat personal emotions** — adds `chat emotion list`, `chat emotion send`, and `chat emotion favorite` for current-user personal favorite emotion listing, sending, and favoriting.

.changes/released/1.0.59-beta.5/minutes-todo-wiki-param-aliases.md

@@ -0,0 +1,5 @@
+---
+category: Added
+---
+
+- **Minutes, DingTalk tasks, and Wiki parameter aliases** — adds reviewed parameter-name normalization, ambiguity guards, and end-to-end payload coverage for the three products.

CHANGELOG.md

@@ -6,6 +6,31 @@ The format is inspired by [Keep a Changelog](https://keepachangelog.com/) and th
 
 ## [Unreleased]
 
+## [1.0.59] - 2026-08-20
+
+This release promotes the sealed `v1.0.59-beta.5` contents to stable.
+
+### Changed
+
+- **Chat personal emotions** — adds commands to list, send, and favorite the current user's personal favorite emotions.
+
+- **Minutes, DingTalk tasks, and Wiki parameter aliases** — adds reviewed parameter-name normalization, ambiguity guards, and end-to-end payload coverage.
+
+- **Shortcut functional workflows** — fixes Drive preview accuracy, AITable write verification and deletion accounting, Wiki feeds, and false-success handling across task, Contact, Minutes, and Wiki operations.
+
+## [1.0.59-beta.5] - 2026-08-20
+
+### Added
+
+- **Chat personal emotions** — adds `chat emotion list`, `chat emotion send`, and `chat emotion favorite` for current-user personal favorite emotion listing, sending, and favoriting.
+
+- **Minutes, DingTalk tasks, and Wiki parameter aliases** — adds reviewed parameter-name normalization, ambiguity guards, and end-to-end payload coverage for the three products.
+
+### Fixed
+
+- **Shortcut functional workflows** (#1050) — fixes truthful Drive push/sync previews, strict AITable write verification and deletion accounting, lossless Wiki feeds, and false-success handling across task, Contact, Minutes, and Wiki operations.
+
+
 ## [1.0.59-beta.4] - 2026-08-20
 
 ### 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.4"
+  version "1.0.59-beta.5"
   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.4/dws-darwin-arm64.tar.gz"
-      sha256 "f788467e9979c70ef210b411ac915b1506ea77ffa496e26b53cfa99650158721"
+      url "https://github.com/DingTalk-Real-AI/dingtalk-workspace-cli/releases/download/v1.0.59-beta.5/dws-darwin-arm64.tar.gz"
+      sha256 "274d56599a8e33ccca86a139424cab95a54ba311d6b643bccb2d3e6608cd16b4"
     else
-      url "https://github.com/DingTalk-Real-AI/dingtalk-workspace-cli/releases/download/v1.0.59-beta.4/dws-darwin-amd64.tar.gz"
-      sha256 "a01988709c0dc99dd5874859eb265ba08a6fda412a7ead8303c68e61d2a8b195"
+      url "https://github.com/DingTalk-Real-AI/dingtalk-workspace-cli/releases/download/v1.0.59-beta.5/dws-darwin-amd64.tar.gz"
+      sha256 "0a0a00a77ca24c102204cd6b7de3de58f406a7e0fad2dd965a4c1fe903c34f39"
     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.4/dws-linux-arm64.tar.gz"
-      sha256 "8e1a993b2137a082a8cc1d9535dfc2d7b3e4399c76d295840f9dc1f15cca7a0d"
+      url "https://github.com/DingTalk-Real-AI/dingtalk-workspace-cli/releases/download/v1.0.59-beta.5/dws-linux-arm64.tar.gz"
+      sha256 "1e7af6393979c2fa433af9207722989749f11ea8e09ff9bcd5696e505d6d7f88"
     else
-      url "https://github.com/DingTalk-Real-AI/dingtalk-workspace-cli/releases/download/v1.0.59-beta.4/dws-linux-amd64.tar.gz"
-      sha256 "26e4cd72cfb96b38ef808863391b81a5c45c3170bca56b5eac457fc601b000c5"
+      url "https://github.com/DingTalk-Real-AI/dingtalk-workspace-cli/releases/download/v1.0.59-beta.5/dws-linux-amd64.tar.gz"
+      sha256 "4896e71a1417acc3d8834fa99f0e81511e020ff952e57c0562cc255e90acec80"
     end
   end
 
   resource "skills" do
-    url "https://github.com/DingTalk-Real-AI/dingtalk-workspace-cli/releases/download/v1.0.59-beta.4/dws-skills.zip"
-    sha256 "a75107bdc14b5476e097842acc92f798301d8ffb59de9ade01f863d166a89435"
+    url "https://github.com/DingTalk-Real-AI/dingtalk-workspace-cli/releases/download/v1.0.59-beta.5/dws-skills.zip"
+    sha256 "11000b9c3566e3b38e3037b6b3069d55c8f50725b3ed9cd67714a7ebb794cd47"
   end
 
   def install

docs/command-index.md

@@ -3,7 +3,7 @@
 Every runtime command the `dws` CLI exposes when loaded with the **pre** environment configuration.
 
 - **Products**: 13
-- **Total commands**: 160
+- **Total commands**: 163
 - **Generated from**: `internal/plugin` command descriptors — the same code path the CLI uses at runtime.
 
 > Auto-generated. Update plugin descriptors in `internal/plugin/`, not this file.
@@ -33,7 +33,7 @@ Every command inherits these flags (documented here once, not repeated per comma
 - [`dws aitable` — AI Tables](#dws-aitable) · 41 commands
 - [`dws attendance` — Attendance](#dws-attendance) · 4 commands
 - [`dws calendar` — Calendar](#dws-calendar) · 14 commands
-- [`dws chat` — Group Chat / IM](#dws-chat) · 23 commands
+- [`dws chat` — Group Chat / IM](#dws-chat) · 26 commands
 - [`dws contact` — Contact Directory](#dws-contact) · 6 commands
 - [`dws devdoc` — Open Platform Docs](#dws-devdoc) · 2 commands
 - [`dws ding` — DING Messages](#dws-ding) · 2 commands
@@ -134,12 +134,15 @@ _Calendar events, participants, meeting rooms, and busy-status queries._
 
 _Group chats, conversations, messages, and robot/webhook integrations._
 
-**23 commands**
+**26 commands**
 
 | Command | Description | When to use |
 |---|---|---|
 | `dws chat bot search` | Search robots (bots) created by the current user by keyword. | When the agent needs to resolve one of its own bots by name to a robot code before sending bot messages. |
 | `dws chat conversation-info` | Retrieve basic metadata for a conversation (single chat or group chat) by conversation ID. | When the agent needs context about a conversation (name, type, member count) before operating on it. |
+| `dws chat emotion favorite` | Add a media ID to the current user's personal favorite emotions. | When the agent needs to save an available mediaId as a reusable personal emotion, optionally preserving source message context. |
+| `dws chat emotion list` | List the current user's personal favorite emotions. | When the agent needs to inspect available personal emotions or resolve an emotionId/mediaId before sending. |
+| `dws chat emotion send` | Send a personal favorite emotion to a group or direct chat as the authenticated user. | When the agent needs to send a known personal emotion mediaId to exactly one group, userId, or openDingTalkId target. |
 | `dws chat group create` | Create a new internal group chat with a set of initial members. | When the agent needs to spin up a dedicated group for a new project, incident, or discussion thread. |
 | `dws chat group members` | List members of a group chat; can also be used against the current user to enumerate their groups' members. | When the agent needs the roster of a group before mentioning, removing, or auditing members. |
 | `dws chat group members add` | Add one or more users to an existing group chat. | When the agent expands a group to include additional participants. |

internal/app/param_alias_e2e_test.go

@@ -68,6 +68,34 @@ func (c *paramAliasCaptureCaller) paramAliasResponseForTool(tool string) string
 		return `{"success":true,"result":[]}`
 	case "list_suggested_event_times":
 		return `{"success":true,"result":{"recommendEventTimes":[]}}`
+	case "list_by_keyword_and_time_range":
+		return `{"success":true,"result":{"itemList":[{"taskUuid":"u1","startTime":1}]}}`
+	case "get_minutes_basic_info":
+		return `{"success":true,"result":{"taskUuid":"u1","title":"Fixture Minutes"}}`
+	case "get_minutes_transcription":
+		return `{"success":true,"result":{"paragraphList":[],"hasNext":false}}`
+	case "create_personal_todo":
+		return `{"success":true,"result":{"taskId":"task-1"}}`
+	case "get_todo_detail":
+		return `{"success":true,"result":{"todoDetailModel":{"taskId":"task-1","subject":"Fixture Todo","isDone":false}}}`
+	case "get_user_todos_in_current_org":
+		return `{"success":true,"result":{"todoCards":[],"hasMore":false}}`
+	case "add_todo_reminder":
+		return `{"success":true}`
+	case "copy_document":
+		return `{"success":true,"nodeId":"copy-1"}`
+	case "move_document", "add_member", "update_member", "remove_member":
+		return `{"success":true}`
+	case "get_document_info":
+		if len(c.calls) > 1 {
+			switch c.calls[len(c.calls)-2].tool {
+			case "copy_document":
+				return `{"success":true,"nodeId":"copy-1","workspaceId":"workspace-1","folderId":"folder-1"}`
+			case "move_document":
+				return `{"success":true,"nodeId":"node-1","workspaceId":"drive-1","folderId":"folder-1"}`
+			}
+		}
+		return `{"success":true,"nodeId":"node-1","workspaceId":"source-1","folderId":"source-folder"}`
 	case "create_calendar_event":
 		return `{"success":true,"result":{"eventId":"event-1"}}`
 	case "update_calendar_event", "delete_calendar_event", "add_calendar_participant", "remove_calendar_participant":

internal/app/param_alias_payload_equivalence_test.go

@@ -232,6 +232,67 @@ var paramAliasCompleteCommands = map[string][]string{
 	"report list":                              {"report", "list", "--start", "2026-03-10T00:00:00+08:00", "--end", "2026-03-10T23:59:59+08:00"},
 }
 
+// paramAliasCandidateCompleteCommands contains complete invocations for the
+// reviewed Minutes/TODO/Wiki joint draft. Keeping candidate-only commands in a
+// separate map lets this test file land before the draft replaces the formal
+// param_concepts.json: inactive candidate templates are ignored, while every
+// command becomes mandatory as soon as one of its reviewed aliases is active.
+var paramAliasCandidateCompleteCommands = map[string][]string{
+	"minutes +detail":             {"minutes", "+detail", "--ids", "u1,u2"},
+	"minutes +latest":             {"minutes", "+latest", "--keyword", "fixture"},
+	"minutes +list-all":           {"minutes", "+list-all", "--limit", "7"},
+	"minutes +record-pause":       {"minutes", "+record-pause", "--id", "u1", "--yes"},
+	"minutes +replace-batch":      {"minutes", "+replace-batch", "--id", "u1", "--pair", "old=>new", "--yes"},
+	"minutes +search":             {"minutes", "+search", "--query", "fixture", "--cursor", "cursor-1"},
+	"minutes +share":              {"minutes", "+share", "--ids", "u1,u2", "--member-uids", "user-1,user-2", "--permission", "view", "--yes"},
+	"minutes +speaker-replace":    {"minutes", "+speaker-replace", "--id", "u1", "--from", "old", "--to", "new", "--target-uid", "user-1", "--yes"},
+	"minutes +summary":            {"minutes", "+summary", "--id", "u1", "--content", "fixture", "--yes"},
+	"minutes +transcript":         {"minutes", "+transcript", "--keyword", "fixture"},
+	"minutes +upload-and-analyze": {"minutes", "+upload-and-analyze", "--resume-id", "u1", "--yes"},
+	"minutes audio-memo list":     {"minutes", "audio-memo", "list", "--max", "7"},
+	"minutes get batch":           {"minutes", "get", "batch", "--ids", "u1,u2"},
+	"minutes hot-word add":        {"minutes", "hot-word", "add", "--words", "DWS,Minutes"},
+	"minutes list all":            {"minutes", "list", "all", "--end", "2026-03-10T23:59:59+08:00"},
+	"minutes list mine":           {"minutes", "list", "mine", "--start", "2026-03-10T00:00:00+08:00"},
+	"minutes replace-text":        {"minutes", "replace-text", "--id", "u1", "--search", "old", "--replace", "new"},
+	"minutes tag query":           {"minutes", "tag", "query", "--tag-id", "tag-1"},
+	"minutes update title":        {"minutes", "update", "title", "--id", "u1", "--title", "Fixture Minutes"},
+	"minutes upload complete":     {"minutes", "upload", "complete", "--session-id", "session-1"},
+	"todo +assign":                {"todo", "+assign", "--task", "Fixture Todo", "--to", "Fixture User", "--yes"},
+	"todo +assign-multi":          {"todo", "+assign-multi", "--task", "Fixture Todo", "--to", "Fixture User,User Two", "--yes"},
+	"todo +comment":               {"todo", "+comment", "--task-id", "task-1", "--content", "fixture comment", "--yes"},
+	"todo +complete":              {"todo", "+complete", "--task-id", "task-1", "--yes"},
+	"todo +create":                {"todo", "+create", "--title", "Fixture Todo", "--executors", "user-1,user-2", "--due", "2026-03-10T18:00:00+08:00", "--yes"},
+	"todo +due-today":             {"todo", "+due-today", "--role-types", "executor"},
+	"todo +get-my-tasks":          {"todo", "+get-my-tasks", "--role-types", "executor", "--priority", "40", "--page", "2", "--size", "7"},
+	"todo +get-related-tasks":     {"todo", "+get-related-tasks", "--role-types", "creator,executor", "--status", "false"},
+	"todo +list-comment":          {"todo", "+list-comment", "--task-id", "task-1", "--page", "2"},
+	"todo +remind":                {"todo", "+remind", "--task", "Fixture Todo", "--at", "2026-03-10T18:00:00+08:00", "--yes"},
+	"todo +reminder":              {"todo", "+reminder", "--task-id", "task-1", "--base-time", "customTime", "--at", "2026-03-10T18:00:00+08:00", "--yes"},
+	"todo +reopen":                {"todo", "+reopen", "--task-id", "task-1", "--yes"},
+	"todo +search":                {"todo", "+search", "--query", "fixture", "--status", "false"},
+	"todo +todo-done":             {"todo", "+todo-done", "--task", "Fixture Todo", "--yes"},
+	"todo +update":                {"todo", "+update", "--task-id", "task-1", "--title", "Fixture Updated Todo", "--yes"},
+	"todo comment add":            {"todo", "comment", "add", "--task-id", "task-1", "--content", "fixture comment", "--yes"},
+	"todo comment list":           {"todo", "comment", "list", "--task-id", "task-1", "--page", "2", "--size", "7"},
+	"todo task add-executor":      {"todo", "task", "add-executor", "--task-id", "task-1", "--executors", "user-1,user-2", "--yes"},
+	"todo task add-participant":   {"todo", "task", "add-participant", "--task-id", "task-1", "--participants", "user-1,user-2", "--yes"},
+	"todo task add-reminder":      {"todo", "task", "add-reminder", "--task-id", "task-1", "--base-time", "customTime", "--reminder-time-stamp", "2026-03-10T18:00:00+08:00", "--yes"},
+	"todo task create":            {"todo", "task", "create", "--title", "Fixture Todo", "--executors", "user-1,user-2", "--due", "2026-03-10T18:00:00+08:00", "--yes"},
+	"todo task create-sub":        {"todo", "task", "create-sub", "--parent-id", "task-parent", "--title", "Fixture Sub Todo", "--executors", "user-1", "--yes"},
+	"todo task done":              {"todo", "task", "done", "--task-id", "task-1", "--status", "true", "--yes"},
+	"todo task get":               {"todo", "task", "get", "--task-id", "task-1"},
+	"todo task list":              {"todo", "task", "list", "--role-types", "executor", "--page", "2", "--size", "7"},
+	"todo task update":            {"todo", "task", "update", "--task-id", "task-1", "--done", "true", "--yes"},
+	"wiki +member-add":            {"wiki", "+member-add", "--workspace", "workspace-1", "--user", "user-1", "--role", "READER", "--yes"},
+	"wiki +member-remove":         {"wiki", "+member-remove", "--workspace", "workspace-1", "--user", "user-1", "--yes"},
+	"wiki +member-update":         {"wiki", "+member-update", "--workspace", "workspace-1", "--user", "user-1", "--role", "EDITOR", "--yes"},
+	"wiki +move":                  {"wiki", "+move", "--workspace", "workspace-1", "--node", "node-1", "--folder", "folder-1", "--yes"},
+	"wiki +move-to-drive":         {"wiki", "+move-to-drive", "--node", "node-1", "--folder", "folder-1", "--yes"},
+	"wiki +node-copy":             {"wiki", "+node-copy", "--workspace", "workspace-1", "--node", "node-1", "--folder", "folder-1", "--yes"},
+	"wiki +node-delete":           {"wiki", "+node-delete", "--workspace", "workspace-1", "--node", "node-1", "--yes"},
+}
+
 // A command can expose more than one mutually exclusive canonical route. In
 // that case the shared command template above cannot contain every canonical
 // flag at once, so select a fixture-specific complete invocation here.
@@ -531,6 +592,20 @@ var paramAliasNewConfirmationCases = []struct {
 	{command: "drive +version-revert", emitted: "version-number", canonical: "version"},
 }
 
+// Candidate confirmation cases become active with the joint draft. One write
+// workflow per product plus TODO's reminder workflow proves semantic aliasing
+// cannot move execution across the shared --yes barrier.
+var paramAliasCandidateConfirmationCases = []struct {
+	command   string
+	emitted   string
+	canonical string
+}{
+	{command: "minutes +record-pause", emitted: "uuid", canonical: "id"},
+	{command: "todo +create", emitted: "deadline", canonical: "due"},
+	{command: "todo +reminder", emitted: "reminder-time-stamp", canonical: "at"},
+	{command: "wiki +node-copy", emitted: "node-id", canonical: "node"},
+}
+
 // paramAliasRepresentativePayloadCases keeps final transport coverage across
 // old concept aliases, command overrides, native compatibility flags, read and
 // write commands, and different products. Every reviewed alias is still
@@ -600,6 +675,30 @@ var paramAliasRepresentativePayloadCases = map[string]bool{
 	paramAliasPayloadCaseKey("report list", "from-date"):                             true, // date-range concept alias
 }
 
+// Candidate representatives exercise the final transport boundary for each
+// Minutes/TODO/Wiki alias family. They are required only when the exact fixture
+// exists in the loaded reviewed table, so the tests are mergeable before the
+// joint draft is promoted to internal/cli/param_concepts.json.
+var paramAliasCandidateRepresentativePayloadCases = map[string]bool{
+	paramAliasPayloadCaseKey("minutes +latest", "query"):               true,
+	paramAliasPayloadCaseKey("minutes +transcript", "query"):           true,
+	paramAliasPayloadCaseKey("minutes get batch", "uuids"):             true,
+	paramAliasPayloadCaseKey("minutes update title", "task-uuid"):      true,
+	paramAliasPayloadCaseKey("minutes upload complete", "upload-id"):   true,
+	paramAliasPayloadCaseKey("todo +create", "deadline"):               true,
+	paramAliasPayloadCaseKey("todo +get-my-tasks", "current-page"):     true,
+	paramAliasPayloadCaseKey("todo +reminder", "reminder-time-stamp"):  true,
+	paramAliasPayloadCaseKey("todo comment add", "text"):               true,
+	paramAliasPayloadCaseKey("todo task add-executor", "executor-ids"): true,
+	paramAliasPayloadCaseKey("todo task get", "todo-id"):               true,
+	paramAliasPayloadCaseKey("todo task update", "status"):             true,
+	paramAliasPayloadCaseKey("wiki +member-add", "user-id"):            true,
+	paramAliasPayloadCaseKey("wiki +member-remove", "uid"):             true,
+	paramAliasPayloadCaseKey("wiki +member-update", "user-id"):         true,
+	paramAliasPayloadCaseKey("wiki +move-to-drive", "node-id"):         true,
+	paramAliasPayloadCaseKey("wiki +node-copy", "node-id"):             true,
+}
+
 // 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
@@ -709,6 +808,7 @@ func TestCrossPlatformCoverageReviewedParamAliasesHaveCompleteTemplatesAndRepres
 	}
 
 	activeCommands := make(map[string]bool)
+	activeFixtureCases := make(map[string]bool)
 	activeCases := 0
 	executedRepresentatives := make(map[string]bool)
 	for _, fixture := range concepts.Fixture {
@@ -717,6 +817,8 @@ func TestCrossPlatformCoverageReviewedParamAliasesHaveCompleteTemplatesAndRepres
 		}
 		activeCommands[fixture.Command] = true
 		activeCases++
+		caseKey := paramAliasPayloadCaseKey(fixture.Command, fixture.Emitted)
+		activeFixtureCases[caseKey] = true
 		complete, ok := paramAliasCompleteCommand(fixture.Command, fixture.Expect)
 		if !ok {
 			t.Errorf("reviewed active fixture %q/%q has no complete-command E2E template", fixture.Command, fixture.Emitted)
@@ -729,8 +831,7 @@ func TestCrossPlatformCoverageReviewedParamAliasesHaveCompleteTemplatesAndRepres
 			continue
 		}
 
-		caseKey := paramAliasPayloadCaseKey(fixture.Command, fixture.Emitted)
-		if !paramAliasRepresentativePayloadCases[caseKey] {
+		if !paramAliasRepresentativePayloadCases[caseKey] && !paramAliasCandidateRepresentativePayloadCases[caseKey] {
 			continue
 		}
 		executedRepresentatives[caseKey] = true
@@ -742,26 +843,43 @@ func TestCrossPlatformCoverageReviewedParamAliasesHaveCompleteTemplatesAndRepres
 	if activeCases == 0 {
 		t.Fatal("reviewed fixture contains no active alias cases")
 	}
+	templateCommands := make(map[string]bool, len(paramAliasCompleteCommands)+len(paramAliasCandidateCompleteCommands))
 	for command := range paramAliasCompleteCommands {
 		if !activeCommands[command] {
 			t.Errorf("complete-command E2E template %q has no active reviewed fixture", command)
 		}
+		templateCommands[command] = true
+	}
+	for command := range paramAliasCandidateCompleteCommands {
+		if activeCommands[command] {
+			templateCommands[command] = true
+		}
 	}
 	for command := range activeCommands {
-		if _, ok := paramAliasCompleteCommands[command]; !ok {
+		if !templateCommands[command] {
 			t.Errorf("active reviewed command %q has no complete-command E2E template", command)
 		}
 	}
-	if len(activeCommands) != len(paramAliasCompleteCommands) {
-		t.Fatalf("complete-command coverage = %d templates for %d active commands (%d active cases)", len(paramAliasCompleteCommands), len(activeCommands), activeCases)
+	if len(activeCommands) != len(templateCommands) {
+		t.Fatalf("complete-command coverage = %d templates for %d active commands (%d active cases)", len(templateCommands), len(activeCommands), activeCases)
 	}
 	for caseKey := range paramAliasRepresentativePayloadCases {
 		if !executedRepresentatives[caseKey] {
 			t.Errorf("representative final-payload case %q has no active reviewed fixture", caseKey)
 		}
 	}
-	if len(executedRepresentatives) != len(paramAliasRepresentativePayloadCases) {
-		t.Fatalf("representative final-payload coverage = %d, want %d", len(executedRepresentatives), len(paramAliasRepresentativePayloadCases))
+	activeRepresentatives := len(paramAliasRepresentativePayloadCases)
+	for caseKey := range paramAliasCandidateRepresentativePayloadCases {
+		if !activeFixtureCases[caseKey] {
+			continue
+		}
+		activeRepresentatives++
+		if !executedRepresentatives[caseKey] {
+			t.Errorf("candidate representative final-payload case %q was not executed", caseKey)
+		}
+	}
+	if len(executedRepresentatives) != activeRepresentatives {
+		t.Fatalf("representative final-payload coverage = %d, want %d", len(executedRepresentatives), activeRepresentatives)
 	}
 }
 
@@ -1091,7 +1209,20 @@ func TestCrossPlatformCoverageNewAITableDeleteDisableAliasesPreserveConfirmation
 }
 
 func TestCrossPlatformCoverageNewParamAliasesCannotBypassConfirmation(t *testing.T) {
-	for _, test := range paramAliasNewConfirmationCases {
+	tests := append([]struct {
+		command   string
+		emitted   string
+		canonical string
+	}{}, paramAliasNewConfirmationCases...)
+	for _, candidate := range paramAliasCandidateConfirmationCases {
+		entry, exists := cli.LookupParamAlias(candidate.command)
+		target, active := entry.ResolveAlias(candidate.emitted)
+		if exists && active && target == candidate.canonical {
+			tests = append(tests, candidate)
+		}
+	}
+
+	for _, test := range tests {
 		test := test
 		t.Run(test.command+"/"+test.emitted, func(t *testing.T) {
 			complete, ok := paramAliasCompleteCommand(test.command, test.canonical)
@@ -1173,6 +1304,10 @@ func paramAliasCompleteCommand(command, canonical string) ([]string, bool) {
 			return variant, true
 		}
 	}
+	if ok {
+		return complete, true
+	}
+	complete, ok = paramAliasCandidateCompleteCommands[command]
 	return complete, ok
 }
 

internal/app/schema_completeness_test.go

@@ -72,6 +72,9 @@ func missingChatCatalogCoveragePaths() []string {
 		"chat clear-messages",
 		"chat clear-red-point",
 		"chat data-auth cross-org",
+		"chat emotion favorite",
+		"chat emotion list",
+		"chat emotion send",
 		"chat group audit-join-validation",
 		"chat group list-all",
 		"chat group list-join-validations",

internal/app/schema_shortcut_contract_test.go

@@ -114,7 +114,7 @@ func TestDeliveryShortcutProgressiveQueriesReturnCompleteContracts(t *testing.T)
 
 	product := executeShortcutSchemaQuery(t, "chat")
 	productPayload, _ := product["product"].(map[string]any)
-	if got, want := int(product["count"].(float64)), 217; got != want {
+	if got, want := int(product["count"].(float64)), 220; got != want {
 		t.Fatalf("schema chat count = %d, want %d", got, want)
 	}
 	summaries := schemaContractObjectSlice(productPayload["tools"])
@@ -140,6 +140,57 @@ func TestDeliveryShortcutProgressiveQueriesReturnCompleteContracts(t *testing.T)
 	assertChatCatalogCompleteLeafContracts(t)
 }
 
+func TestChatPersonalEmotionSchemaDeclaresUnpinnedIMAdapter(t *testing.T) {
+	for _, tc := range []struct {
+		cliPath string
+		params  map[string]string
+	}{
+		{
+			cliPath: "chat emotion list",
+		},
+		{
+			cliPath: "chat emotion send",
+			params: map[string]string{
+				"media-id":         "mediaId",
+				"emotion-id":       "emotionId",
+				"group":            "openConversationId",
+				"open-dingtalk-id": "receiverOpenDingTalkId",
+				"idempotency-key":  "uuid",
+			},
+		},
+		{
+			cliPath: "chat emotion favorite",
+			params: map[string]string{
+				"media-id":               "mediaId",
+				"name":                   "name",
+				"source-conversation-id": "sourceConversationId",
+				"source-message-id":      "sourceMessageId",
+			},
+		},
+	} {
+		t.Run(tc.cliPath, func(t *testing.T) {
+			leaf := executeShortcutSchemaQuery(t, "--cli-path", tc.cliPath)
+			if got := schemaContractString(leaf["interface_mode"]); got != "composite" {
+				t.Fatalf("%s interface_mode = %q, want composite", tc.cliPath, got)
+			}
+			reason := schemaContractString(leaf["interface_reason"])
+			if !strings.Contains(reason, "Reviewed unpinned remote adapter") {
+				t.Fatalf("%s interface_reason = %q", tc.cliPath, reason)
+			}
+			parameters := schemaContractMap(leaf["parameters"])
+			for name, want := range tc.params {
+				parameter := parameters[name]
+				if parameter == nil {
+					t.Fatalf("%s missing --%s parameter: %#v", tc.cliPath, name, parameters)
+				}
+				if got := schemaContractString(parameter["property"]); got != want {
+					t.Fatalf("%s --%s property = %q, want %q", tc.cliPath, name, got, want)
+				}
+			}
+		})
+	}
+}
+
 func TestCrossPlatformCoverageAITableTableBootstrapPublishesResultContract(t *testing.T) {
 	leaf := executeShortcutSchemaQuery(t, "--cli-path", "aitable +table-bootstrap")
 	result, _ := leaf["result"].(map[string]any)

internal/helpers/chat.go

@@ -10569,7 +10569,7 @@ pl_PL, sv_SE, fi_FI, cs_CZ, ar_SA, tl_PH, he_IL, nl_NL, lo_LA, it_IT`,
 	chatCategoryCmd.AddCommand(chatCategoryCreateSmartCmd)
 	chatMessageCmd.AddCommand(chatMessageListDirectCmd, chatMessageSearchCommonCmd, chatMessageCombineForwardCmd, chatMessageForwardTopicCmd, chatMessageSetPinCmd, chatMessageUnsetPinCmd, chatMessageListPinCmd, chatMessageAddFavoriteCmd, chatMessageRemoveFavoriteCmd, chatMessageListFavoritesCmd, chatMessageSetTopMsgCmd, chatMessageUnsetTopMsgCmd, chatMessageListEmotionRepliesCmd)
 
-	root.AddCommand(chatChmodCmd, chatDataAuthCmd, chatGroupCmd, chatSearchCmd, chatSearchCommonCmd, chatMessageCmd, chatFileCmd, newChatMediaGroup(), chatBotCmd, chatMessageListTopConversationsCmd, chatConversationInfoCmd, chatCategoryCmd, chatGroupRoleCmd, chatMuteCmd, chatSetTopCmd, chatGroupMuteCmd, chatGroupMuteMemberCmd, chatHideCmd, chatMuteAtAllCmd, chatMuteRedEnvelopeCmd, chatMarkUnreadCmd, chatClearRedPointCmd, chatClearAllRedPointCmd, chatListAllConversationsCmd, chatClearMessagesCmd, chatMarkReadCmd, chatTextCmd, newChatToolbarCommand())
+	root.AddCommand(chatChmodCmd, chatDataAuthCmd, chatGroupCmd, chatSearchCmd, chatSearchCommonCmd, chatMessageCmd, chatFileCmd, newChatMediaGroup(), chatBotCmd, chatMessageListTopConversationsCmd, chatConversationInfoCmd, chatCategoryCmd, chatGroupRoleCmd, chatMuteCmd, chatSetTopCmd, chatGroupMuteCmd, chatGroupMuteMemberCmd, chatHideCmd, chatMuteAtAllCmd, chatMuteRedEnvelopeCmd, chatMarkUnreadCmd, chatClearRedPointCmd, chatClearAllRedPointCmd, chatListAllConversationsCmd, chatClearMessagesCmd, chatMarkReadCmd, chatTextCmd, newChatToolbarCommand(), newChatEmotionCommand())
 
 	// Keep the v1.0.56 command surface recognizable while directing callers to
 	// the supported nested commands. The chat root's "im" alias makes these

internal/helpers/chat_personal_emotion.go

@@ -0,0 +1,253 @@
+package helpers
+
+import (
+	"fmt"
+	"strings"
+
+	"github.com/DingTalk-Real-AI/dingtalk-workspace-cli/internal/cli"
+	"github.com/DingTalk-Real-AI/dingtalk-workspace-cli/internal/corecmd/contract"
+	"github.com/DingTalk-Real-AI/dingtalk-workspace-cli/internal/shortcut/targetresolver"
+	"github.com/spf13/cobra"
+)
+
+const personalEmotionUnpinnedReason = "Reviewed unpinned remote adapter: this executable CLI wrapper calls a remote helper that is absent from the pinned MCP metadata snapshot; no single pinned semantically equivalent interface_ref can represent the command."
+
+func newChatEmotionCommand() *cobra.Command {
+	cmd := &cobra.Command{
+		Use:   "emotion",
+		Short: "个人收藏表情",
+		Long:  "查询、发送和新增当前用户的个人收藏表情。",
+		RunE:  groupRunE,
+	}
+	cmd.AddCommand(
+		newChatEmotionListCommand(),
+		newChatEmotionSendCommand(),
+		newChatEmotionFavoriteCommand(),
+	)
+	return cmd
+}
+
+func newChatEmotionListCommand() *cobra.Command {
+	cmd := &cobra.Command{
+		Use:   "list",
+		Short: "列出个人收藏表情",
+		RunE: func(cmd *cobra.Command, args []string) error {
+			return callMCPToolOnServer("im", "list_personal_emotions", map[string]any{})
+		},
+	}
+	DeclareLeafMetadata(cmd, LeafSpec{
+		Safety: contract.SafetySpec{
+			Effect: "read", Risk: "low", Confirmation: "not_required", Idempotency: "idempotent",
+		},
+		Contract: LeafContract{
+			Identity: contract.ToolIdentitySpec{
+				ProductID: "chat", Name: "list_personal_emotions",
+				CanonicalPath: "chat.list_personal_emotions", CLIPath: "chat emotion list", PrimaryCLIPath: "chat emotion list",
+			},
+			Description: "列出当前用户的个人收藏表情",
+			Interface: &contract.InterfaceSpec{
+				Mode: "composite", Availability: "available", Reason: personalEmotionUnpinnedReason,
+			},
+			Selection: contract.SelectionSpec{
+				AgentSummary: "列出当前用户的个人收藏表情",
+				UseWhen:      []string{"需要查看当前用户已收藏的表情、获取 emotionId 或 mediaId 时"},
+				AvoidWhen:    []string{"查询消息 reaction 使用 chat message list-emotion-replies"},
+				Examples:     []string{"dws chat emotion list --format json"},
+			},
+		},
+	})
+	return cmd
+}
+
+func newChatEmotionSendCommand() *cobra.Command {
+	cmd := &cobra.Command{
+		Use:   "send",
+		Short: "发送个人收藏表情",
+		Long: `发送当前用户的个人收藏表情。
+
+⚠️ 重要：该接口会真实发送表情到目标会话，不可用于测试或试探性调用。调用前必须确认表情媒体 ID 和接收对象无误。`,
+		Example: `  dws chat emotion send --media-id <mediaId> --group <openConversationId>
+  dws chat emotion send --media-id <mediaId> --emotion-id <emotionId> --user <userId>
+  dws chat emotion send --media-id <mediaId> --open-dingtalk-id <openDingTalkId> --uuid <idempotencyKey>`,
+		RunE: func(cmd *cobra.Command, args []string) error {
+			mediaID, _ := cmd.Flags().GetString("media-id")
+			if strings.TrimSpace(mediaID) == "" {
+				return fmt.Errorf("--media-id is required")
+			}
+			target, err := personalEmotionSendTarget(cmd)
+			if err != nil {
+				return err
+			}
+			payload := map[string]any{"mediaId": strings.TrimSpace(mediaID)}
+			emotionID, _ := cmd.Flags().GetString("emotion-id")
+			if strings.TrimSpace(emotionID) != "" {
+				payload["emotionId"] = strings.TrimSpace(emotionID)
+			}
+			for key, value := range target {
+				payload[key] = value
+			}
+			if uuid := strings.TrimSpace(flagOrFallback(cmd, "uuid", "idempotency-key")); uuid != "" {
+				payload["uuid"] = uuid
+			}
+			return callMCPToolOnServer("im", "send_personal_emotion", payload)
+		},
+	}
+	cmd.Flags().String("media-id", "", "表情媒体 ID (必填)")
+	cmd.Flags().String("emotion-id", "", "表情 ID")
+	cmd.Flags().String("conversation-id", "", "群聊 openConversationId")
+	cmd.Flags().String("group", "", "群聊 openConversationId（--conversation-id 别名）")
+	cmd.Flags().String("user", "", "单聊接收人 userId；CLI 会解析为 openDingTalkId")
+	cmd.Flags().String("open-dingtalk-id", "", "单聊接收人 openDingTalkId")
+	cmd.Flags().String("uuid", "", "幂等键")
+	cmd.Flags().String("idempotency-key", "", "幂等键（--uuid 别名）")
+	cmd.MarkFlagsMutuallyExclusive("conversation-id", "group")
+	cli.AnnotateRuntimeConstraints(cmd, cli.RuntimeSchemaConstraints{
+		MutuallyExclusive: [][]string{{"conversation-id", "group", "user", "open-dingtalk-id"}},
+		RequireOneOf:      [][]string{{"conversation-id", "group", "user", "open-dingtalk-id"}},
+	})
+	DeclareLeafMetadata(cmd, LeafSpec{
+		Safety: contract.SafetySpec{
+			Effect: "write", Risk: "medium", Confirmation: "not_required", Idempotency: "unknown",
+		},
+		Contract: personalEmotionSendContract(),
+	})
+	return cmd
+}
+
+func newChatEmotionFavoriteCommand() *cobra.Command {
+	cmd := &cobra.Command{
+		Use:   "favorite",
+		Short: "新增个人收藏表情",
+		Example: `  dws chat emotion favorite --media-id <mediaId> --name "赞"
+  dws chat emotion favorite --media-id <mediaId> --source-conversation-id <cid> --source-message-id <mid>`,
+		RunE: func(cmd *cobra.Command, args []string) error {
+			mediaID, _ := cmd.Flags().GetString("media-id")
+			if strings.TrimSpace(mediaID) == "" {
+				return fmt.Errorf("--media-id is required")
+			}
+			sourceConversationID, _ := cmd.Flags().GetString("source-conversation-id")
+			sourceMessageID, _ := cmd.Flags().GetString("source-message-id")
+			if err := validatePersonalEmotionSourcePair(sourceConversationID, sourceMessageID); err != nil {
+				return err
+			}
+			payload := map[string]any{"mediaId": strings.TrimSpace(mediaID)}
+			name, _ := cmd.Flags().GetString("name")
+			if strings.TrimSpace(name) != "" {
+				payload["name"] = strings.TrimSpace(name)
+			}
+			if strings.TrimSpace(sourceConversationID) != "" {
+				payload["sourceConversationId"] = strings.TrimSpace(sourceConversationID)
+				payload["sourceMessageId"] = strings.TrimSpace(sourceMessageID)
+			}
+			return callMCPToolOnServer("im", "favorite_personal_emotion", payload)
+		},
+	}
+	cmd.Flags().String("media-id", "", "待收藏 mediaId (必填)")
+	cmd.Flags().String("name", "", "表情名称")
+	cmd.Flags().String("source-conversation-id", "", "来源会话 ID；需与 --source-message-id 成对指定")
+	cmd.Flags().String("source-message-id", "", "来源消息 ID；需与 --source-conversation-id 成对指定")
+	DeclareLeafMetadata(cmd, LeafSpec{
+		Safety: contract.SafetySpec{
+			Effect: "write", Risk: "medium", Confirmation: "not_required", Idempotency: "unknown",
+		},
+		Contract: personalEmotionFavoriteContract(),
+	})
+	return cmd
+}
+
+func personalEmotionSendTarget(cmd *cobra.Command) (map[string]any, error) {
+	groupID := strings.TrimSpace(flagOrFallback(cmd, "conversation-id", "group"))
+	userID, _ := cmd.Flags().GetString("user")
+	openDingTalkID, _ := cmd.Flags().GetString("open-dingtalk-id")
+	userID = strings.TrimSpace(userID)
+	openDingTalkID = strings.TrimSpace(openDingTalkID)
+	specified := 0
+	for _, value := range []string{groupID, userID, openDingTalkID} {
+		if value != "" {
+			specified++
+		}
+	}
+	if specified != 1 {
+		return nil, fmt.Errorf("--conversation-id, --user or --open-dingtalk-id is required; specify exactly one")
+	}
+	if groupID != "" {
+		return map[string]any{"openConversationId": groupID}, nil
+	}
+	if openDingTalkID != "" {
+		if err := targetresolver.ValidateExplicitOpenDingTalkID("--open-dingtalk-id", openDingTalkID); err != nil {
+			return nil, err
+		}
+		return map[string]any{"receiverOpenDingTalkId": openDingTalkID}, nil
+	}
+	if isOpenDingTalkID(userID) {
+		return map[string]any{"receiverOpenDingTalkId": userID}, nil
+	}
+	resolved, err := resolveOpenDingTalkID(cmd.Context(), userID)
+	if err != nil {
+		return nil, fmt.Errorf("cannot resolve --user %q to openDingTalkId: %w; pass --open-dingtalk-id instead", userID, err)
+	}
+	return map[string]any{"receiverOpenDingTalkId": resolved}, nil
+}
+
+func validatePersonalEmotionSourcePair(sourceConversationID, sourceMessageID string) error {
+	hasConversation := strings.TrimSpace(sourceConversationID) != ""
+	hasMessage := strings.TrimSpace(sourceMessageID) != ""
+	if hasConversation != hasMessage {
+		return fmt.Errorf("--source-conversation-id and --source-message-id must be specified together")
+	}
+	return nil
+}
+
+func personalEmotionSendContract() LeafContract {
+	return LeafContract{
+		Identity: contract.ToolIdentitySpec{
+			ProductID: "chat", Name: "send_personal_emotion",
+			CanonicalPath: "chat.send_personal_emotion", CLIPath: "chat emotion send", PrimaryCLIPath: "chat emotion send",
+		},
+		Description: "以当前用户身份向群聊或单聊发送个人收藏表情",
+		Interface: &contract.InterfaceSpec{
+			Mode: "composite", Availability: "available", Reason: personalEmotionUnpinnedReason,
+		},
+		Selection: contract.SelectionSpec{
+			AgentSummary: "以当前用户身份发送个人收藏表情",
+			UseWhen:      []string{"用户明确要求发送个人收藏表情，且已提供 mediaId 或 emotionId 时"},
+			AvoidWhen:    []string{"发送普通文本、Markdown 或文件时使用 chat message send"},
+			Examples:     []string{"dws chat emotion send --media-id <mediaId> --group <openConversationId> --uuid <idempotencyKey>"},
+		},
+		Parameters: []contract.ParamDecl{
+			{Name: "media-id", Property: "mediaId", Required: boolPtr(true)},
+			{Name: "emotion-id", Property: "emotionId", Required: boolPtr(false)},
+			{Name: "conversation-id", Property: "openConversationId", Required: boolPtr(false)},
+			{Name: "group", Property: "openConversationId", Required: boolPtr(false)},
+			{Name: "user", Property: "receiverOpenDingTalkId", Required: boolPtr(false)},
+			{Name: "open-dingtalk-id", Property: "receiverOpenDingTalkId", Required: boolPtr(false)},
+			{Name: "uuid", Property: "uuid", Required: boolPtr(false)},
+			{Name: "idempotency-key", Property: "uuid", Required: boolPtr(false)},
+		},
+	}
+}
+
+func personalEmotionFavoriteContract() LeafContract {
+	return LeafContract{
+		Identity: contract.ToolIdentitySpec{
+			ProductID: "chat", Name: "favorite_personal_emotion",
+			CanonicalPath: "chat.favorite_personal_emotion", CLIPath: "chat emotion favorite", PrimaryCLIPath: "chat emotion favorite",
+		},
+		Description: "将 mediaId 新增到当前用户的个人收藏表情",
+		Interface: &contract.InterfaceSpec{
+			Mode: "composite", Availability: "available", Reason: personalEmotionUnpinnedReason,
+		},
+		Selection: contract.SelectionSpec{
+			AgentSummary: "新增当前用户的个人收藏表情",
+			UseWhen:      []string{"用户要把一个 mediaId 收藏为个人表情时"},
+			AvoidWhen:    []string{"收藏消息使用 chat message add-favorite"},
+			Examples:     []string{"dws chat emotion favorite --media-id <mediaId> --name \"赞\""},
+		},
+		Parameters: []contract.ParamDecl{
+			{Name: "media-id", Property: "mediaId", Required: boolPtr(true)},
+			{Name: "name", Property: "name", Required: boolPtr(false)},
+			{Name: "source-conversation-id", Property: "sourceConversationId", Required: boolPtr(false), RequiredWhen: "source-message-id is provided"},
+			{Name: "source-message-id", Property: "sourceMessageId", Required: boolPtr(false), RequiredWhen: "source-conversation-id is provided"},
+		},
+	}
+}

internal/helpers/chat_personal_emotion_test.go

@@ -0,0 +1,254 @@
+package helpers
+
+import (
+	"context"
+	"io"
+	"reflect"
+	"strings"
+	"testing"
+
+	"github.com/DingTalk-Real-AI/dingtalk-workspace-cli/pkg/edition"
+)
+
+type personalEmotionCall struct {
+	server string
+	tool   string
+	args   map[string]any
+}
+
+type personalEmotionCaller struct {
+	calls []personalEmotionCall
+}
+
+func (c *personalEmotionCaller) CallTool(_ context.Context, server, tool string, args map[string]any) (*edition.ToolResult, error) {
+	copied := make(map[string]any, len(args))
+	for key, value := range args {
+		copied[key] = value
+	}
+	c.calls = append(c.calls, personalEmotionCall{server: server, tool: tool, args: copied})
+	if server == "contact" && tool == "get_user_info_by_user_ids" {
+		return textToolResult(`{"result":[{"userId":"u1","openDingTalkId":"` + helperCurrentDOpenID2 + `"}]}`), nil
+	}
+	return textToolResult(`{"ok":true}`), nil
+}
+
+func (*personalEmotionCaller) Format() string { return "json" }
+func (*personalEmotionCaller) DryRun() bool   { return false }
+func (*personalEmotionCaller) Fields() string { return "" }
+func (*personalEmotionCaller) JQ() string     { return "" }
+
+func executePersonalEmotionCommand(t *testing.T, caller *personalEmotionCaller, args ...string) error {
+	t.Helper()
+	installHelpersCoreDeps(t, caller)
+	deps.Out.w = io.Discard
+	deps.Out.errW = io.Discard
+	root := newChatCommand()
+	root.SilenceErrors = true
+	root.SilenceUsage = true
+	root.SetOut(io.Discard)
+	root.SetErr(io.Discard)
+	root.SetArgs(args)
+	return root.ExecuteContext(context.Background())
+}
+
+func requirePersonalEmotionCall(t *testing.T, caller *personalEmotionCaller, tool string, want map[string]any) {
+	t.Helper()
+	if len(caller.calls) != 1 {
+		t.Fatalf("calls = %d, want 1: %+v", len(caller.calls), caller.calls)
+	}
+	call := caller.calls[0]
+	if call.server != "im" || call.tool != tool {
+		t.Fatalf("tool call = %s/%s, want im/%s", call.server, call.tool, tool)
+	}
+	if !reflect.DeepEqual(call.args, want) {
+		t.Fatalf("args = %#v, want %#v", call.args, want)
+	}
+}
+
+func TestChatEmotionListCallsIMToolWithoutBusinessArgs(t *testing.T) {
+	// TC-001: list 无业务参数，当前用户身份由 MCP server 注入。
+	caller := &personalEmotionCaller{}
+	if err := executePersonalEmotionCommand(t, caller, "emotion", "list"); err != nil {
+		t.Fatalf("chat emotion list returned error: %v", err)
+	}
+	requirePersonalEmotionCall(t, caller, "list_personal_emotions", map[string]any{})
+}
+
+func TestChatEmotionSendMapsGroupTargetAndIdempotency(t *testing.T) {
+	// TC-002: 群聊目标映射为 openConversationId，uuid 与表情字段按 MCP 字段透传。
+	caller := &personalEmotionCaller{}
+	err := executePersonalEmotionCommand(t, caller,
+		"emotion", "send",
+		"--media-id", "@media",
+		"--emotion-id", "emotion123",
+		"--group", "cid123",
+		"--idempotency-key", "idem-001",
+	)
+	if err != nil {
+		t.Fatalf("chat emotion send returned error: %v", err)
+	}
+	requirePersonalEmotionCall(t, caller, "send_personal_emotion", map[string]any{
+		"mediaId":            "@media",
+		"emotionId":          "emotion123",
+		"openConversationId": "cid123",
+		"uuid":               "idem-001",
+	})
+}
+
+func TestChatEmotionSendMapsOpenDingTalkTarget(t *testing.T) {
+	// TC-003: 已知 openDingTalkId 时直传 receiverOpenDingTalkId，不做外部解析。
+	caller := &personalEmotionCaller{}
+	err := executePersonalEmotionCommand(t, caller,
+		"emotion", "send",
+		"--media-id", "@media",
+		"--open-dingtalk-id", helperCurrentDOpenID,
+	)
+	if err != nil {
+		t.Fatalf("chat emotion send returned error: %v", err)
+	}
+	requirePersonalEmotionCall(t, caller, "send_personal_emotion", map[string]any{
+		"mediaId":                "@media",
+		"receiverOpenDingTalkId": helperCurrentDOpenID,
+	})
+}
+
+func TestChatEmotionSendTreatsOpenDingTalkIDPassedAsUserAsResolvedTarget(t *testing.T) {
+	// TC-004: --user 收到 openDingTalkId 形态时保持 chat message send 的兼容语义。
+	caller := &personalEmotionCaller{}
+	err := executePersonalEmotionCommand(t, caller,
+		"emotion", "send",
+		"--media-id", "@media",
+		"--user", helperCurrentDOpenID,
+	)
+	if err != nil {
+		t.Fatalf("chat emotion send returned error: %v", err)
+	}
+	requirePersonalEmotionCall(t, caller, "send_personal_emotion", map[string]any{
+		"mediaId":                "@media",
+		"receiverOpenDingTalkId": helperCurrentDOpenID,
+	})
+}
+
+func TestChatEmotionSendResolvesUserIDTarget(t *testing.T) {
+	// TC-004b: --user 收到普通 userId 时先解析为 openDingTalkId，再发送个人收藏表情。
+	caller := &personalEmotionCaller{}
+	err := executePersonalEmotionCommand(t, caller,
+		"emotion", "send",
+		"--media-id", "@media",
+		"--user", "u1",
+	)
+	if err != nil {
+		t.Fatalf("chat emotion send returned error: %v", err)
+	}
+	if len(caller.calls) != 2 {
+		t.Fatalf("calls = %d, want contact resolve then send: %+v", len(caller.calls), caller.calls)
+	}
+	resolveCall := caller.calls[0]
+	if resolveCall.server != "contact" || resolveCall.tool != "get_user_info_by_user_ids" {
+		t.Fatalf("resolve call = %s/%s, want contact/get_user_info_by_user_ids", resolveCall.server, resolveCall.tool)
+	}
+	sendCall := caller.calls[1]
+	if sendCall.server != "im" || sendCall.tool != "send_personal_emotion" {
+		t.Fatalf("send call = %s/%s, want im/send_personal_emotion", sendCall.server, sendCall.tool)
+	}
+	want := map[string]any{
+		"mediaId":                "@media",
+		"receiverOpenDingTalkId": helperCurrentDOpenID2,
+	}
+	if !reflect.DeepEqual(sendCall.args, want) {
+		t.Fatalf("send args = %#v, want %#v", sendCall.args, want)
+	}
+}
+
+func TestChatEmotionSendRejectsInvalidTargets(t *testing.T) {
+	tests := []struct {
+		name    string
+		args    []string
+		wantErr string
+	}{
+		{
+			name:    "missing target",
+			args:    []string{"emotion", "send", "--media-id", "@media"},
+			wantErr: "specify exactly one",
+		},
+		{
+			name:    "multiple targets",
+			args:    []string{"emotion", "send", "--media-id", "@media", "--group", "cid", "--open-dingtalk-id", helperCurrentDOpenID},
+			wantErr: "specify exactly one",
+		},
+		{
+			name:    "missing media",
+			args:    []string{"emotion", "send", "--group", "cid"},
+			wantErr: "--media-id is required",
+		},
+	}
+	for _, tc := range tests {
+		t.Run(tc.name, func(t *testing.T) {
+			caller := &personalEmotionCaller{}
+			err := executePersonalEmotionCommand(t, caller, tc.args...)
+			if err == nil || !strings.Contains(err.Error(), tc.wantErr) {
+				t.Fatalf("error = %v, want containing %q", err, tc.wantErr)
+			}
+			if len(caller.calls) != 0 {
+				t.Fatalf("invalid command reached MCP: %+v", caller.calls)
+			}
+		})
+	}
+}
+
+func TestChatEmotionFavoriteMapsOptionalSourcePair(t *testing.T) {
+	// TC-005: 收藏来源字段成对出现时透传为 sourceConversationId/sourceMessageId。
+	caller := &personalEmotionCaller{}
+	err := executePersonalEmotionCommand(t, caller,
+		"emotion", "favorite",
+		"--media-id", "@media",
+		"--name", "赞",
+		"--source-conversation-id", "cid123",
+		"--source-message-id", "msg123",
+	)
+	if err != nil {
+		t.Fatalf("chat emotion favorite returned error: %v", err)
+	}
+	requirePersonalEmotionCall(t, caller, "favorite_personal_emotion", map[string]any{
+		"mediaId":              "@media",
+		"name":                 "赞",
+		"sourceConversationId": "cid123",
+		"sourceMessageId":      "msg123",
+	})
+}
+
+func TestChatEmotionFavoriteRejectsMissingRequiredOrUnpairedSource(t *testing.T) {
+	tests := []struct {
+		name    string
+		args    []string
+		wantErr string
+	}{
+		{
+			name:    "missing media",
+			args:    []string{"emotion", "favorite", "--name", "赞"},
+			wantErr: "--media-id is required",
+		},
+		{
+			name:    "source conversation only",
+			args:    []string{"emotion", "favorite", "--media-id", "@media", "--source-conversation-id", "cid123"},
+			wantErr: "must be specified together",
+		},
+		{
+			name:    "source message only",
+			args:    []string{"emotion", "favorite", "--media-id", "@media", "--source-message-id", "msg123"},
+			wantErr: "must be specified together",
+		},
+	}
+	for _, tc := range tests {
+		t.Run(tc.name, func(t *testing.T) {
+			caller := &personalEmotionCaller{}
+			err := executePersonalEmotionCommand(t, caller, tc.args...)
+			if err == nil || !strings.Contains(err.Error(), tc.wantErr) {
+				t.Fatalf("error = %v, want containing %q", err, tc.wantErr)
+			}
+			if len(caller.calls) != 0 {
+				t.Fatalf("invalid command reached MCP: %+v", caller.calls)
+			}
+		})
+	}
+}

skills/mono/references/products/chat.md

@@ -1212,6 +1212,51 @@ Flags:
       --msg-ids string   消息 ID 列表，逗号分隔 (必填)
 ```
 
+### emotion (个人收藏表情)
+
+#### 列出个人收藏表情
+```
+Usage:
+  dws chat emotion list [flags]
+Example:
+  dws chat emotion list --format json
+```
+当前用户身份由 MCP server 注入，无业务参数。
+
+#### 发送个人收藏表情
+```
+Usage:
+  dws chat emotion send [flags]
+Example:
+  dws chat emotion send --media-id <mediaId> --group <openConversationId>
+  dws chat emotion send --media-id <mediaId> --emotion-id <emotionId> --user <userId>
+  dws chat emotion send --media-id <mediaId> --open-dingtalk-id <openDingTalkId> --uuid <idempotencyKey>
+Flags:
+      --media-id string            表情媒体 ID (必填)
+      --emotion-id string          表情 ID (可选)
+      --conversation-id string     群聊 openConversationId（与 --group 等价）
+      --group string               群聊 openConversationId
+      --user string                单聊接收人 userId，CLI 会解析为 openDingTalkId
+      --open-dingtalk-id string    单聊接收人 openDingTalkId
+      --uuid string                幂等键
+      --idempotency-key string     幂等键（--uuid 等价）
+```
+目标参数 `--conversation-id/--group`、`--user`、`--open-dingtalk-id` 三选一。该命令会真实发送表情，执行前必须确认目标和 mediaId。
+
+#### 新增个人收藏表情
+```
+Usage:
+  dws chat emotion favorite [flags]
+Example:
+  dws chat emotion favorite --media-id <mediaId> --name "赞"
+  dws chat emotion favorite --media-id <mediaId> --source-conversation-id <cid> --source-message-id <mid>
+Flags:
+      --media-id string                  待收藏 mediaId (必填)
+      --name string                      表情名称
+      --source-conversation-id string    来源会话 ID，需与 --source-message-id 成对指定
+      --source-message-id string         来源消息 ID，需与 --source-conversation-id 成对指定
+```
+
 ### list-top-conversations (置顶会话)
 
 #### 拉取置顶会话列表
@@ -2129,6 +2174,9 @@ Flags:
 用户说"群公告/发布公告/修改公告/查看公告/定时公告" → `chat group notice create/edit/get/list`
 用户说"批量查消息/按ID查消息/根据消息ID查" → `chat message list-by-ids`
 用户说"批量查消息回复/表情回复/文字回复/消息回应列表" → `chat message list-emotion-replies`
+用户说"查看个人收藏表情/列出我的收藏表情" → `chat emotion list`
+用户说"发送个人收藏表情/发表情包" → `chat emotion send`
+用户说"收藏表情/新增个人收藏表情" → `chat emotion favorite`
 用户说"emoji回应/表情回应/给消息加表情" → `chat message add-emoji`
 用户说"取消emoji回应/移除表情回应" → `chat message remove-emoji`
 用户说"文字表情回应/添加文字表情" → `chat message add-text-emotion`
@@ -2200,6 +2248,7 @@ Flags:
 - `chat message list-emotion-replies` — 批量拉取消息的表情回复和文字回复
 - `chat message add-text-emotion` / `update-text-emotion` / `remove-text-emotion` — 对消息添加、原地更新或移除文字表情回应
 - `chat message create-text-emotion` — 创建文字表情模板，返回 emotionId 供 add-text-emotion 使用
+- `chat emotion list` / `send` / `favorite` — 当前用户个人收藏表情列表、发送和新增；不同于消息 reaction/文字回应
 - `chat category list` — 获取用户自定义会话分组列表
 - `chat category list-conversations` — 拉取指定分组下的会话列表
 - `chat category create-smart` — 创建智能会话分组（可指定群名称关键词和成员作为匹配规则）

skills/multi/dingtalk-chat/SKILL.md

@@ -47,6 +47,7 @@ metadata:
 | 查看指定群成员（用户/机器人） | `dws chat +chat-members-list --group <群名或ID>` | 唯一解析并全量读取 |
 | 获取群邀请链接 | `dws chat +chat-invite-url --group <群名或ID>` | 多候选时停止 |
 | 查看群机器人 | `dws chat +chat-bots --group <群名或ID>` | 返回稳定 `bots[]` |
+| 个人收藏表情列表/发送/收藏 | `dws chat emotion list/send/favorite` | 约束见 leaf Schema |
 | 修改群名称 | `dws chat group rename --id <openConversationId> --name <新名称>` | 只知群名时先用 `+chat-search --query <群名>` 唯一解析 ID；不猜 `+chat-rename` |
 | 查看指定群内 @我的消息 | `dws chat +at-me --group <群名> --page-all` | 检查 `complete`；空结果仍返回数组 |
 | 查看全部会话 | `dws chat +conversation-list --page-all` | 检查 `complete` / `failures` |

skills/multi/dingtalk-chat/references/chat.md

@@ -57,8 +57,10 @@
 | `message add-emoji` / `remove-emoji` | 默认 emoji reaction |
 | `message create-text-emotion` / `add-text-emotion` / `update-text-emotion` / `remove-text-emotion` | 文字表情 |
 | `message list-emotion-replies` | 批量 reaction/文字回应 |
+| `emotion list` / `send` / `favorite` | 当前用户个人收藏表情列表、发送和新增 |
 
 Favorite、消息 Pin、消息 Top 与会话 Top 是四种对象，不能互换。
+个人收藏表情与消息 reaction/文字回应不同；发送收藏表情使用 `chat emotion send`，给已有消息贴表情使用 `chat message add-emoji` 或 `chat message add-text-emotion`。
 
 ## 群与成员底层能力
 

skills/multi/dingtalk-doc/SKILL.md

@@ -1,6 +1,6 @@
 ---
 name: dingtalk-doc
-description: 钉钉在线文字文档（adoc）本体及其内容的操作：查找、创建、读取、文档信息、编辑、块、评论、附件与媒体、白板卡片、导入、导出(docx/markdown/pdf)、版本、模板、权限、分享及Markdown/JSONML写入。不包括：文档空间与钉盘的文件管理（归 dingtalk-drive，doc 同名原子命令已弃用）、知识库空间与节点管理（归 dingtalk-wiki）、原生 .md 文件读写（归 dingtalk-misc）、电子表格 axls（归 dingtalk-misc）、AI 表格 able（归 dingtalk-aitable）。命令前缀：dws doc。
+description: 钉钉在线文字文档（adoc）的查找创建、读写、文档信息、白板卡片、附件、评论、媒体、导入导出、版本、模板、权限分享；本地文件转在线文档或协作编辑归 dingtalk-doc。原文件/文件夹归 dingtalk-drive，知识库归 dingtalk-wiki，原生 .md/axls 归 dingtalk-misc，able 归 dingtalk-aitable。前缀：dws doc。
 metadata:
   cli_version: ">=0.2.14"
   category: product
@@ -35,21 +35,23 @@ metadata:
 
 ## Golden Route
 
-选择最小入口。ID/URL 直用；只有标题时先搜索，唯一候选后执行，禁止默认第一项。
+走最小入口：ID/URL 直用；标题先搜索，唯一命中再执行。成本顺序：稳定 ID → shortcut → 局部读 → 精确写；禁以产品 Schema、全文或原子命令起步。
 
 | 用户意图 | 唯一推荐入口 | 关键边界 |
 |---|---|---|
 | 按标题或主题定位文档 | `dws doc +search --query <关键词>` | 检查候选类型与分页；需要正文时再用真实 `nodeId` 执行 `+fetch` |
-| 已知 ID/URL 读取正文或局部内容 | `dws doc +fetch --node <ID或URL>` | `--scope keyword/section/range` 可减少无关正文；非 adoc 切换对应产品 |
+| 最近访问文档 | `dws doc +search`（省略 `--query`） | `--limit` 为每页量，`--max-items` 为总上限；完整集合才加 `--page-all` 并检查 `complete` |
+| 已知 ID/URL 读取正文或局部内容 | `dws doc +fetch --node <ID或URL>` | 具体术语直用 `keyword`；章节先 `outline` 再 `section`；只有整篇任务才用默认 `full` |
 | 聚合查看信息、权限、版本、媒体或评论 | `dws doc +inspect --node <ID或URL>` | 仅打开任务所需的 `--include-*`，不要默认全取 |
-| 新建在线文字文档并写入内容 | `dws doc +create --name <标题> --content @<相对文件>` | 长或复杂 Markdown 使用 `@file`/stdin；根据结果验证真实 `nodeId` |
-| 追加、覆盖或精确编辑 block | `dws doc +update --node <ID或URL> --command <动作>` | `--expected-revision` 仅用于 JSONML 原子覆盖；先消费 Schema confirmation |
+| 新建在线文字文档并写入内容 | `dws doc +create --name <标题> --content @<相对文件>` | 先在本地完成正文；Runtime 负责 Markdown 分片和回读，禁止 Agent 自行拆成多次远程写入 |
+| 追加、覆盖或精确编辑 block | `dws doc +update --node <ID或URL> --command <动作>` | 唯一文本直接 `str_replace`；章节/block 先局部取 ID 再精准修改；整篇覆盖才用 overwrite |
 | 重要内容更新且需要恢复点 | `dws doc +checkpoint-update` | 自动保存版本、更新并回读；检查 `steps` 和 `compensation` |
 | 版本操作 | `dws doc +version-save --node` / `dws doc +version-list --node` / `dws doc +version-revert --node --version` | 快照/列表/回滚 |
 | 导出为 docx/markdown/pdf | `dws doc +export --export-format <格式>` | 格式必须显式指定；普通文件下载切 `dingtalk-drive` |
-| 导入本地文件为在线对象 | `dws doc +import --file <相对路径>` | folder/workspace 可选，缺省导入默认根；纯上传切 `dingtalk-drive` |
-| 浏览模板 | `dws doc +template-list [--source MY\|PUBLIC]` | 没有名称/关键词时使用；不调用 search/create |
-| 搜索模板 | `dws doc +template-search --query <名称或关键词>` | 有查询词才使用；零或多候选停止 |
+| 本地文件转在线文档 | `dws doc +import --file <相对路径>` | “传上去/放进文件夹”不改变路由；在线改、协作编辑或转换用 import，仅保留原文件切 `dingtalk-drive` |
+| 封面/背景 | `+resource-update/+resource-delete`；`+background-update/+background-delete` | 写后 `+inspect --include-style`；禁查 Catalog |
+| 浏览模板 | `dws doc +template-list [--source MY\|PUBLIC]` | “我的/我这边”只查 MY；明确公开才查 PUBLIC；“有哪些/全部”翻页至完整 |
+| 搜索模板 | `dws doc +template-search --query <名称或关键词>` | 来源可选 MY/PUBLIC；零命中停止，禁止拿无关模板替代；多候选消歧 |
 | 从模板创建 | `dws doc +create-from-template --template-id <唯一ID>` | 已有唯一 templateId 才创建；不重复 list/search |
 | 创建评论或聚合待处理评论 | `dws doc +comment-create [--selection]` / `+review` | 划词统一用 `+comment-create`；后续操作使用真实 `commentKey` |
 | 添加/调整/移除协作者权限 | `dws doc +access-grant/+access-change/+access-revoke` | 先读取现有权限；姓名歧义或 profile 不一致时禁止写入 |
@@ -58,18 +60,20 @@ metadata:
 
 ## 关键结果语义
 
-- 保留真实 `nodeId`、URL、资源类型和容器；标题不能替代稳定 ID。
-- 按状态恢复：`partial_success` 只补未完成步骤；`unknown` 先回读且不重试写入；`retryable` 仅限明确未开始；权限、参数、认证失败停止。
+- 保留真实 `nodeId`、URL、资源类型和容器；同轮创建后直接复用返回 ID，禁止再按标题或钉盘搜索。
+- 先消费完整回执；Runtime 已分片/回读时不重复读取。只有局部验收、`partial_success` 或 commit-unknown 才再 `+fetch`。
+- 状态恢复：`partial_success` 只补未完成步骤；`unknown` 先回读、不重试写；`retryable` 仅限明确未开始；权限/参数/认证失败停止。
 - 仅在结果明确且关键内容回读匹配后报告写入完成。
-- 搜索、列表和批量结果必须检查 `complete`、`hasMore`、continuation 和失败项；单页结果不得表述为完整集合。
+- 搜索/列表检查 `complete`、`hasMore`、cursor 和失败项。“有哪些/列出/全部”逐页至完整；只有示例/前 N 条可提前停止并声明范围。
+- `+import` 已含上传、转换和轮询；成功检查 `success=true`、`taskId`、`documentUrl`。超时或中断保留 `taskId`、查询原任务，禁止重导。
 - 导出/下载仅用工作目录相对路径，默认不覆盖并原子落盘。
 
 ## 参数与安全边界
 
-- `@file` 统一协议：已有或临时文件先暂存到 cwd 后传 `@相对路径`；单次生成文本优先 `--content -` (stdin)；禁止绝对路径和 `..`。
+- `@file`：已有或临时文件先暂存到 cwd；传 `@相对路径`，禁绝对路径和 `..`。
 - `doc +update` 的动作由 `--command` 指定；block 操作的 ID 必须来自 `+fetch --detail with-ids` 或真实 block 列表。
-- Schema 门禁：已知只读路由且参数明确时直接执行；写命令或 selection/参数不清、准备 Help 时，本轮仅查一次 `dws schema --cli-path "doc +<leaf>" --compact --fields use_when,avoid_when,parameters,constraints,confirmation -f json`；禁用产品级/`--all`。
-- 执行前消费其中的 `confirmation`：`user_required` 且用户已确认同一目标/动作/参数才加 `--yes`，否则预览/询问；禁止靠失败探测门禁。
+- Schema 门禁：不确定时仅查一次精确 leaf：`--fields use_when,avoid_when,parameters,constraints,confirmation`；禁用产品级/`--all`。准备 Help 时，本轮仅查一次。
+- 消费本页或精确 Schema 的 `confirmation`：`user_required` 且原请求/预授权已确认目标、动作、参数时，首调即加 `--yes`；否则预览/询问；禁止靠失败探测门禁。
 - JSONML 顶层必须是单个非空元素；禁止 `[[...]]` 元素数组包裹。
 
 ## 按需加载
@@ -80,25 +84,24 @@ Golden Route 已给出命令且参数足够时，禁止读取 reference。其余
 |---|---|
 | 低频/无 shortcut 意图消歧 | [intent-guide.md](references/intent-guide.md) / [doc.md](references/doc.md) 对应章节 |
 | 分页、`partial_success`、`status=unknown` 或恢复 | [contracts.md](references/contracts.md) |
-| 复杂 JSONML/非默认操作 | [create](references/doc/doc-create.md) / [read](references/doc/doc-read.md) / [update](references/doc/doc-update.md) |
-| block/划词评论/媒体高级参数 | [block](references/doc/doc-block.md) / [comment](references/doc/doc-comment.md) / [media](references/doc/doc-media.md) |
+| 复杂 JSONML、长文或局部精准读写 | [create](references/doc/doc-create.md) / [read](references/doc/doc-read.md) / [update](references/doc/doc-update.md) |
+| block/划词评论/媒体/封面/背景高级参数 | [block](references/doc/doc-block.md) / [comment](references/doc/doc-comment.md) / [media](references/doc/doc-media.md) |
 | 导出/导入失败恢复 | [export](references/doc/doc-export.md) / [import](references/doc/doc-import.md) |
 
-`+create`、`+fetch`、`+update` append/overwrite、`+export`、`+import` 禁止读取 reference；禁止预加载或连读。
+常规 `+create`、`+fetch`、`+update` append/overwrite、`+export`、`+import` 禁止读取 reference；需策略时仅读上表一项，禁预加载/连读。
 
 ## 错误最短路径
 
 1. 零命中、多候选、类型不明或分页不完整：停止后续写入，展示候选或 continuation；禁止默认第一项。
 2. Help 不参与选路；按上方门禁先消费一次精确 leaf Schema。只有真实 `unknown flag`/契约漂移后才查一次 leaf Help；`unknown command` 只查一次 shortcut 清单，禁止试探后缀和 `dws doc --help | grep/head`。
 3. `REVISION_CONFLICT`：重新读取当前 revision，展示差异；未经用户确认不得改成无 revision 覆盖。
-4. `partial_success`：保留已完成步骤，优先执行补偿或从 checkpoint 继续，不重放成功步骤。
-5. `doc_write_commit_unknown`：先回读文档确认状态；禁止自动重试创建或追加。
-6. 认证、权限或 profile 错误：只读取 `dingtalk-shared` 对应 reference；不要尝试同义底层命令绕过。
-7. 导出或媒体失败：保留 `jobId/resourceId/nodeId` 后停止；禁止 `curl`、安装依赖、Python 文档库或手写 HTTP 兜底。
+4. `doc_write_commit_unknown`：先回读；禁止自动重试创建或追加。
+5. 认证、权限或 profile 错误：只读 `dingtalk-shared` 对应 reference，不用底层命令绕过。
+6. 导出/媒体失败：保留稳定 ID 后停止；禁止网络请求、安装依赖或本地文档库兜底。
 
 ## 跨产品边界
 
-- 普通文件、目录、纯上传下载、节点存储权限 → `dingtalk-drive`
+- 普通文件/目录/纯上传下载/节点权限 → `dingtalk-drive`；保留原文件用原子命令 `dws drive upload --workspace`，在线转换用 `doc +import --workspace`
 - 知识库空间、节点层级和成员管理 → `dingtalk-wiki`
 - 原生 `.md` 文件读取和编辑 → `dingtalk-misc`
 - `axls` / `able` → 对应电子表格或多维表 Skill

skills/multi/dingtalk-doc/references/contracts.md

@@ -13,10 +13,12 @@
 | `success` | 所有计划步骤完成，且需要验证的内容已经回读 | 可以向用户报告完成 |
 | `partial_success` | 已发生部分副作用，后续步骤失败 | 检查 `steps` 与 `compensation`，不得重放成功步骤 |
 | `unknown` | 请求已发出但无法确定服务端是否提交 | 先回读目标；创建和追加禁止自动重试 |
-| `retryable` | 服务端明确业务执行尚未开始，且允许重试 | 遵循 `retry_after`，最多有界重试一次 |
-| `failed` | 已确认没有完成目标动作 | 根据 `retryable` 和 `suggestedActions` 决定是否重试 |
+| `retryable` | 服务端明确业务执行尚未开始，且允许重试 | 遵循 `retry_after_seconds`，最多有界重试一次 |
+| `failed` | 已确认没有完成目标动作 | 根据 `retryable`、`actions` 和 details 决定是否重试 |
 
-`ok=true` 或进程退出码为零不能替代 `status/verified`。写回执应保留 operation、target、steps、warnings、failures、compensation 和 verification 摘要。
+进程退出码为零不能替代业务证据。采用 `doc.operation.v1` 的 shortcut 回执固定提供 `contractVersion/ok/status/complete/operation/steps/data/warnings/compensation`；`target/failures/verification` 只有实际操作返回时才能消费，不是通用字段。`+import` 使用现有导入回执，成功时检查 `success=true`、`taskId`、`documentUrl`、`documentName` 与 `documentType`，不要要求不存在的 `status/steps`。业务 `status` 不应与框架外层 `outcome` 混为一谈。
+
+稳定返回 ID/任务 ID 只用于定位目标、恢复流程或继续查询，不能单独证明写操作成功。成功证据优先级从高到低为：与该操作匹配的明确服务端成功终态 → 契约要求的写后读回匹配 → 仅 transport/退出码；异步任务必须使用返回的任务 ID 查询到成功终态。只有成功终态成立且所有必要回读均匹配时才报告成功；`partial_success`、`unknown`、未完成任务或仅返回资源/任务 ID 都不得报告完成。Runtime 已给出充分回读证据时，不为“再确认一次”重复请求。
 
 ## 分页与完整性
 
@@ -24,14 +26,14 @@
 
 - `complete=true`：已证明覆盖请求范围；
 - `hasMore=true`：仍有后续页，必须保留有效 continuation；
-- `truncated=true`：因 maxPages/maxItems/timeout 等边界停止；
-- `failures[]`：已返回部分数据但某一页或后处理失败。
+- `truncated=true`：成功返回因 `max_pages` 或 `max_items` 边界停止，并通过 `stopReason` 说明原因；
+- 后续页请求失败、cursor 缺失/停滞/循环时返回 typed error；部分结果位于 `details.items`，并保留 `details.status=partial_success`、`complete=false`、`reason/page/nextCursor/count`。不要期待成功回执的空 `failures[]` 承载这类错误，也不要把 timeout 写成 `truncated`。
 
-只有 `complete=true` 且没有未处理失败时，才能把结果描述为完整集合。
+只有 `complete=true` 且没有未处理失败时，才能把结果描述为完整集合。用户问“有哪些”“全部”或要求“列出结果”时，使用返回的 cursor/continuation 继续取页直至完整；只有用户明确要示例、前 N 条或接受部分结果时才可提前停止，并说明已覆盖范围。分页应复用原查询与过滤条件，不得换命令或放宽关键词。
 
 ## 错误
 
-结构化错误至少区分 validation、not_found、ambiguous、type_mismatch、revision_conflict、confirmation_required、permission_denied、partial_success 和 commit_unknown，并提供 failure stage、retryable 和 suggested actions。权限、认证和参数错误直接进入 `failed`；写请求只有明确 `execution_started=false` 才能进入 `retryable`，其余传输异常进入 `unknown`。`retryable=false` 表示自动重放不安全，不代表用户检查状态后永远不能重新发起。
+结构化错误至少区分 validation、not_found、ambiguous、type_mismatch、revision_conflict、confirmation_required、permission_denied、partial_success 和 commit_unknown，并提供 failure stage、retryable 与已有的 `actions`/details。权限、认证和参数错误直接进入 `failed`；写请求只有明确 `execution_started=false` 才能进入 `retryable`，其余传输异常进入 `unknown`。`retryable=false` 表示自动重放不安全，不代表用户检查状态后永远不能重新发起。不要发明 `suggestedAction` 或顶层 `nextCommand` 字段。
 
 ## 安全落盘
 

skills/multi/dingtalk-doc/references/doc.md

@@ -26,10 +26,14 @@
 dws doc +template-search --query "周报" --source PUBLIC --format json
 ```
 
+来源按用户原话守门：“我的模板/我这边”只查 `MY`，明确“公开/钉钉模板库”才查 `PUBLIC`；不得为了凑结果跨来源扩展。未指定来源时保持默认 `MY`。
+
 - `selection.status=resolved`：取唯一候选的 `templateId`。
-- `selection.status=not_found`：报告零命中后停止。
+- `selection.status=not_found`：报告零命中后停止；不得改用语义不相干的热门模板，更不得擅自创建文档。
 - `selection.status=selection_required`：展示候选并要求用户选择，禁止默认第一项。
 
+若返回 `hasMore=true`，沿原 query/source 使用 cursor 继续搜索；只有服务端返回完整结果后才能判断零命中或完整候选集。
+
 选定后只创建一次：
 
 ```bash

skills/multi/dingtalk-doc/references/doc/doc-create.md

@@ -12,7 +12,7 @@ dws doc +create --name "<文档名>" --content @body.json --doc-format jsonml --
 
 - 统一输入协议：已有或临时文件先暂存到当前工作目录后传 `@相对文件`；单次生成文本可用 `--content -` 从 stdin 读取。
 - `@file` 禁止绝对路径和 `..` 逃逸；不要直接引用宿主临时目录。
-- `--name` 是文档名称，不等于正文中的 heading。用户明确要求“正文一级标题”时，正文仍须包含对应 H1。
+- `--name` 是文档名称；默认不要再在正文开头重复同名 H1，只有用户明确要求正文一级标题时才保留。
 - JSONML 顶层必须是数组；仅在确实需要富结构时加载 JSONML cookbook。
 
 `+create` 负责创建、长 Markdown 分片和最终回读验证。正常成功结果至少包含：
@@ -35,6 +35,18 @@ dws doc +create --name "<文档名>" --content @body.json --doc-format jsonml --
 - `status=unknown`：服务端可能已经提交；先定位并读取文档，禁止自动重试。
 - 没有真实 `nodeId` 或写回执时，禁止声称“已创建”。
 
+## 创作与执行策略
+
+采用 Plan → Execute → Observe → Iterate，将循环放在本地内容与定点修正上，不把长文拆成一串远程创建/追加：
+
+1. **Plan**：明确受众、目的、范围和结构；正文由一个主上下文串行维护，避免按章节并行生成造成重复、矛盾和语气漂移。
+2. **Draft**：短内容可直接传；多行、长文或含特殊字符时先形成 cwd 内相对文件。用户未要求富结构时优先 Markdown；只有样式/引用/嵌套结构确有必要时才用 JSONML。
+3. **Execute**：只调用一次 `+create`。DWS Runtime 会分片长 Markdown、记录每步并回读验证；Agent 不采用“先骨架、再逐节远程插入”流程，以减少网络次数、顺序错误和 commit-unknown 面。
+4. **Observe**：先检查回执中的 `nodeId`、`verified`、`steps` 和失败状态。回执已证明完整时不重复拉全文；需要内容质量验收时，只用 `+fetch --scope section/keyword` 读取待检查部分。
+5. **Iterate**：后续修正复用同一 `nodeId`，按 [`doc-update.md`](doc-update.md) 做最小 block/文本修改，禁止重新创建整篇。
+
+交付前检查标题是否重复、段落是否连贯、编号是否统一；只有真实行列数据才使用表格，富组件服务于理解而不是装饰。明确字数要求时应在写入前完成本地统计，不能凭模型估算宣称达标。
+
 ## 高级通道
 
 只有 shortcut 未公开所需的底层参数或需要原始响应时，才读取精确 leaf Schema 后使用 `dws doc create`。不要因为熟悉旧参数就默认退回原子命令，也不要使用已删除的 Python 创建脚本。

skills/multi/dingtalk-doc/references/doc/doc-import.md

@@ -8,14 +8,15 @@ dws doc +import --file ./report.docx --folder <FOLDER_ID> --format json
 dws doc +import --file ./notes.md --workspace <WORKSPACE_ID> --name "会议纪要" --format json
 ```
 
-`+import` 一次完成创建会话、上传、确认转换和轮询。支持 `doc/docx/xls/xlsx/md/txt/xmind/mark`，文件大小上限 20MB。
+`+import` 一次完成创建会话、上传、确认转换和终态轮询。支持 `doc/docx/xls/xlsx/md/txt/xmind/mark`，文件大小上限 20MB。转换成功回执包含 `success=true`、`taskId`、`documentUrl`、`documentName` 和 `documentType`，不包含 `status` 或 `steps`；成功返回即表示本次内部轮询已到终态。超时或中断时保留错误中的 `taskId`，只查询原任务。
 
 ## 本地文件边界
 
 - `--file` 只接受当前工作目录内已存在的相对路径；禁止绝对路径、`..` 或符号链接逃逸。
-- `--folder` 与 `--workspace` 都是可选位置；都不传时导入默认根目录，两者都有时优先 `--folder`。
+- `--folder` 与 `--workspace` 都是可选位置；都不传时导入默认根目录。通常只提供一个；当前 Runtime 在两者同时出现时会一并传给服务端，未声明客户端优先级，不要依赖 `--folder` 覆盖 `--workspace`。
 - CLI 负责上传和格式转换。不要先用 Python/Office 库解析文件，不要安装本地依赖来伪造在线导入结果，也不要手写 HTTP 上传。
 - 白名单外格式（如 HTML/PDF）自动改走原文件上传，返回 `fallback=upload`、`converted=false`；不得报告成已经转换为可编辑在线文档。
+- “在线改/协作编辑/转在线文档”属于导入转换；“存着/归档/保留原文件/不要转换”属于 `dingtalk-drive` 纯上传。目标为文档空间时，纯上传使用 `drive upload --workspace <WORKSPACE_ID>`，不要因容器叫“文档空间”就误报为在线文档。
 
 ## 失败处理
 

skills/multi/dingtalk-doc/references/doc/doc-media.md

@@ -16,10 +16,22 @@ dws doc +media-download --node <DOC_ID> --resource-id <RESOURCE_ID> --output ./d
 dws doc +media-preview --node <DOC_ID> --resource-id <RESOURCE_ID> --format json
 ```
 
+## 封面与背景 Shortcut
+
+```bash
+dws doc +resource-update --node <DOC_ID> --file ./cover.png --format json
+dws doc +resource-download --node <DOC_ID> --output ./cover.png --format json
+dws doc +resource-delete --node <DOC_ID> --format json
+dws doc +background-update --node <DOC_ID> --color "#E8F2FE" --format json
+dws doc +background-delete --node <DOC_ID> --format json
+```
+
+封面不是正文媒体 block；背景仅接受 `#RRGGBB` 纯色。设置/清除后用一次 `+inspect --include-style` 验证，禁止为这些已知能力查询 shortcut Catalog。
+
 ## 稳定 ID 与结果
 
 - `resourceId`、`blockId`、`nodeId` 必须来自真实 media/block 返回，不能从标题或本地文件名猜测。
-- 插入后保留 `resourceId` 和插入结果；下载必须检查 `localPath`、`sizeBytes > 0`。
+- 插入成功回执在 `data.blockId` 返回已回读验证的媒体块 ID；后续定位只复用该 `blockId`。回执不提供相邻空块字段，不得据此猜测或自动删除其他块。插入回执成功后禁止重传媒体。下载必须检查 `localPath`、`sizeBytes > 0`。
 - 下载输出只接受工作目录内相对路径，默认 no-clobber。
 - 删除源文件是独立的破坏性本地操作，不属于媒体下载；只有用户明确要求且下载验证成功后才能执行。
 

skills/multi/dingtalk-doc/references/doc/doc-read.md

@@ -24,10 +24,24 @@ dws doc +fetch --node <DOC_ID> --scope keyword --keyword "风险|结论" --conte
 
 只有需要块 ID、revision 或 JSONML 保真结构时才提高 `--detail`；先读取最小必要范围，避免把整篇大文档放入上下文。
 
+整篇读取使用默认 scope 或 `--scope full`；`full` 不是关键词，禁止写成 `--keyword full`。
+
+## 最小读取漏斗
+
+按用户已经给出的线索选最短路径，不先拉全文：
+
+1. 已知稳定 ID/URL，且任务确实涉及整篇：直接默认 `full + simple`，一次返回 Markdown。
+2. 用户给出具体术语、错误码或同义词：直接 `keyword`；`foo|bar` 是 OR，`context-before/after` 是字符数。该模式会缩小 Agent 输出与 token，但当前 Runtime 仍需取得正文后做本地投影，不把它误述为减少上游传输。
+3. 用户指向某章但没有 block ID：先 `outline --detail with-ids`，再用返回的真实标题 ID 执行 `section`。通常两次小结果比反复处理整篇更稳定。
+4. 已知起止 block：直接 `range`；只找表格、图片等结构时用 `tags`。
+5. 只有确需全文总结、全局一致性检查或整篇保真改写时，才读取完整内容。
+
+`simple` 用于阅读；`with-ids` 用于定位下一次写操作；`full` 只用于必须保留样式、引用或 JSONML 结构的编辑。局部结果只能证明所选范围，不得据此声称已检查整篇。
+
 ## 后续路由
 
 - 读后修改：把稳定的 `nodeId` 交给 [`doc-update.md`](doc-update.md) 的 `+update` 或 `+checkpoint-update`。
 - 附件/图片：先用 [`doc-media.md`](doc-media.md) 的 `+media-list` 取得稳定 `resourceId`，再下载；不要复用正文中的临时签名 URL。
 - 富结构专家编辑：确实需要原始 JSONML 或 shortcut 未公开的参数时，先读取精确 leaf Schema，再使用原子 `doc read`/`doc block`。
 
-禁止把原子 `doc read` 当作默认入口，也不要在读取后无条件整篇回写。筛选结果只用于读取，不能把虚拟 fragment 容器整体写回文档。
+禁止把原子 `doc read` 当作默认入口，也不要在读取后无条件整篇回写。筛选结果只用于读取，不能把虚拟 fragment 容器整体写回文档；同一次任务里已拿到稳定 `nodeId` 后，后续调用直接复用，避免再次按标题搜索。

skills/multi/dingtalk-doc/references/doc/doc-update.md

@@ -4,6 +4,8 @@
 
 普通追加、覆盖和 block 编辑统一使用 `+update`：
 
+`--command` 只接受下列枚举值，不接受 JSON、自然语言或拼接子命令；动作参数必须分别传给 `--content/--old/--new/--block-id/--after-block-id`。
+
 ```bash
 dws doc +update --node <DOC_ID> --command append --content "补充说明" --format json
 dws doc +update --node <DOC_ID> --command append --content @append.md --format json
@@ -38,9 +40,29 @@ block ID 必须来自 `+fetch --detail with-ids` 或真实 block 列表，禁止
 
 `--expected-revision` 只允许 `--command overwrite --doc-format jsonml`。Markdown、append 和 block 接口没有服务端原子 revision 契约，禁止用写前读取模拟乐观锁。
 
+## 最小改写决策
+
+| 已知条件 | 推荐路径 | 成本与成功率理由 |
+|---|---|---|
+| 用户明确要求在末尾追加 | 直接 `append` | 不为找末尾先拉全文；需要语气衔接时只读末节 |
+| 已知唯一旧文本与新文本 | 直接 `str_replace --old --new` | 省掉 block 解析；旧文本不唯一时 Runtime 必须失败，不放宽匹配 |
+| 指定章节但没有 block ID | `+fetch outline` → `+fetch section --detail with-ids` → block 动作 | 两个小读取换取稳定锚点，避免全文 token 和误改相邻章节 |
+| 已知真实 block ID | 直接 `block_replace/delete/insert_after` | 最小副作用；不改无关 block |
+| 多处富结构保真修改 | `+fetch --detail full` 后定点 JSONML 更新 | 保留图片、附件、引用、表格和样式；不要从 Markdown 有损重建 |
+| 整篇重要覆盖 | `+checkpoint-update --mode overwrite` | 自动保存恢复点、执行并回读；普通 overwrite 只用于明确不需恢复点的场景 |
+
+同一篇文档的正文由一个主上下文串行维护：Plan（确定最小变更）→ Execute（一次写）→ Observe（先消费回执）→ Iterate（只修未达标部分）。不要按章节并行写同一文档，也不要每次迭代都重新读取全文或 Schema。
+
+## Block ID 生命周期与保真
+
+- `block_replace` 成功后 Runtime 使用同一 `blockId` 回读验证，该 ID 可继续作为锚点；验证失败时先局部 `+fetch` 核对现状。`block_delete` 成功后旧 ID 失效，不得继续复用。
+- `block_insert_after` / `block_copy_insert_after` 后，原锚点通常仍可识别，但新 block 的 ID 必须来自真实返回或局部回读，禁止按顺序猜测。
+- `str_replace` 的简单行内替换通常不要求重新取 ID；若后续依赖块结构，仍以局部回读为准。
+- 从 Markdown 读取后覆盖整篇可能丢失图片、附件、@人/@文档、评论锚点、表格样式和嵌套块。只改局部时使用 block 手术；确需整篇保真改写时使用 `full` JSONML，并以 `--expected-revision` 防止覆盖并发修改。
+
 ## 确认与验证
 
-- 执行前从精确 leaf Schema 消费 `confirmation`；需要确认时先 `--dry-run`/询问，用户确认同一参数后追加 `--yes`，禁止先失败探测门禁。
+- `+update` 与 `+checkpoint-update` 当前都要求用户确认。目标、动作、内容范围或参数变化后必须重新确认；只有发现本地文档与 live leaf 漂移时才查一次精确 Schema，禁止每次写入都重复发现。
 - `+update` 已负责回读验证。除非结果为 `unknown` 或任务需要额外结构验收，不要再做一次全篇读取。
 - `doc_write_verification_failed` 表示写入已经发生，必须先读取现状；禁止直接重复写入。
 - `partial_success` 只恢复未完成步骤，不能重放成功步骤。