Compile reviewed Agent Schema into bind-time Go declarations so catalog
tools stamp contract_final without changing execution bodies.
Co-authored-by: Cursor <cursoragent@cursor.com>
Align the write-confirmation UX with lark-cli: ConfirmRisk (and the
shortcut confirmRisk) now print the yes/no prompt only when stdin is a
real terminal (ioctl-level check via go-isatty; a char-device stat would
misclassify `< /dev/null`). Non-interactive callers get a clean
structured confirmation_required error on stderr. Piped answers are
still honored for humans/scripts; --yes/--dry-run remain the sanctioned
non-interactive paths.
skills/mono: add the recognition + retry protocol for agents —
identify confirmation_required via error.reason, show action and params,
retry the original command with --yes only after explicit user consent,
never silently append --yes or treat it as a transient error.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Co-authored-by: Cursor <cursoragent@cursor.com>
Treat EOF/closed stdin as confirmation_required instead of an
interactive decline so agent/CI no longer get exit 0 for writes that
never ran. Align cmdcore.ConfirmRisk the same way.
Co-authored-by: Cursor <cursoragent@cursor.com>
The single Dispatch hook assumed every command is one MCP call, so the
Shortcut projection could only ever describe a command, never run it. Split
dispatch into Invoke (assembled toolArgs) and Orchestrate (multi-step), keep
RunE as the escape hatch, and reject specs that declare anything other than
exactly one at construction time. Ctx gives both hooks the same typed flag
accessors so orchestration no longer reaches for framework-specific plumbing.
Catalog output stays byte-identical.
Adversarial review confirmed the Phase 1 extraction is a verbatim move (no BLOCKERs) and that cmdcore.BoolFlag is equivalent to both original readers. All fixes below are in the Phase 2 additions.
Correctness: a CommandSpec declaring neither RunE nor Dispatch no longer runs the whole pipeline (write-confirmation prompt included) and silently exits 0 — it now fails with a typed internal error, which also defuses the FromShortcut trap. FromShortcut no longer double-renders the 参数约束 section (shortcutLongHelp already appends it and NewCommand appends ConstraintHelp again): it now maps intent prose only. Flag usage keeps mount()'s flagHelp decoration (必填/可选值) so projected help matches the live shortcut, and the constraint Flags slice is copied instead of aliasing the shortcut registry.
Honesty: the FromShortcut doc block now lists every dropped semantic (Required Changed-vs-effective-value divergence, typed bool/int/slice defaults, Enum, Hidden, Tips to Example, custom constraint, required/enum runtime-schema annotations). cmdcore's package doc no longer claims catalog drift proves runtime behavior — drift covers the build-time projection, unit tests cover the runtime pipeline. leaf.go's stale Phase 1 header updated. Panic messages say command not leaf; doc comments lead with the exported names.
Tests: new mount-equivalence test compares the projected command's flag set/types/usage and rendered Long against live mount(s) — the test that would have caught both bugs above; new root-to-child test exercises the inherited/root-persistent --yes/--dry-run lookup that the leaf-local helper never reached; the nil-dispatch test is inverted to assert the error. docs/architecture.md documents internal/cmdcore and how it differs from internal/cobracmd.
Verified: drift ok (840 tools, identical hashes), make policy pass, skill-command-integrity ok (1033 paths), cmdcore self-coverage 100%, CI-style changed-code coverage 100% (278 statements), full helpers/cmdcore/shortcut suites green.
Introduce cmdcore.CommandSpec as the single typed leaf definition and
cmdcore.NewCommand as the one orchestration path (flags → constraint decl
checks → Runtime Schema projection → constraint help → PostMount → RunE
escape / generated RunE{required → constraints → Validate → BuildArgs →
ConfirmRisk → Dispatch}). Dispatch becomes a spec property, not a
separate framework.
helpers.NewLeafCommand now delegates to cmdcore.NewCommand(FromLeafSpec),
so every LeafSpec command — including all 27 devapp leaves — flows through
the unified spec. The MCP dispatch (Call / callMCPToolOnServer /
callMCPTool) is captured in the FromLeafSpec closure.
internal/shortcut/adapter.go adds FromShortcut, the typed seam mapping a
Shortcut's shared base (flags of every kind, known constraints, risk,
help identity) into a CommandSpec. It is intentionally NOT wired into the
live mount() path: Shortcut's multi-step Execute, decline-returns-nil
semantics, and Flag.Enum/Hidden/custom-constraint extras are not modeled
by cmdcore yet, so the 376 shipped shortcuts stay byte-identical. Live
wiring is deferred to Phase 3, gated by shortcut-list + schema equivalence.
Commands are provably unaffected: check-generated-drift ok (840 tools,
identical hashes), `dws schema --all` and `shortcut list` unchanged, full
helpers/cmdcore/shortcut suites green, changed-code coverage 100%.
Two independent shortcut correctness fixes surfaced by the audit:
- Name→ID resolution (chat +dm / +broadcast / … via the shared resolver) dropped
every search_contact_by_key_word row with an empty userId. External /
cross-org contacts arrive with only an openDingTalkId, so they were silently
discarded — making resolution report a real person as missing, or collapse to
the wrong single match when an in-org namesake existed. Keep any row with at
least one usable identity (userId or openDingTalkId) and fall the display name
back through nick/showName/flowerName/staffName/userName.
- chat +messages-resource-url required --message-id with no alias, so an agent
copying the message list's openMessageId/msgId output field hit "unknown
flag". Accept --msg-id / --open-message-id as aliases (declared via an
at-least-one constraint since a shortcut's Required check only sees the
primary flag name), mirroring the earlier chat message download-media fix.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Publishes all 210 public built-in shortcuts as reviewed Agent-visible
leaf tools across 16 product groups, with stable canonical identities,
executable +shortcut CLI paths, parameter and cross-parameter
constraints, selection guidance, interface metadata, and runtime-aligned
safety/confirmation semantics. Catalog grows from 603 to 813 tools.
- minutes: drop the minutesId/minutes_id candidate from the taskUuid mapping.
minutesId is the minutes document id, a different identifier from the
recording taskUuid that +record-pause/resume/stop consume via --id, so
substituting it would feed record control a wrong id. The backend list
already returns taskUuid; the guard test now asserts taskUuid/task_uuid.
- Rewrite the guard-test failure messages and fixture data in English to match
the repository convention (only the two assertions that match the
production Chinese validation string are kept).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Several read shortcuts returned an empty list with exit 0 and no error
envelope even though the underlying MCP tool returned data, so agents misread
"no data" and made wrong decisions.
Root causes:
- Container key mismatch: the resolver probed the wrong key —
processCodeList / values / wikiSpaces / itemList / groupList / recentItems /
emailAccounts / deptUserList / labelUserList / roles / report_list, plus
get_org_labels grouped labels[] needing a descend.
- Item fields nested under a VO wrapper, not unwrapped: shiftVO / entityVO /
userInfo.
- Param exceeded a backend limit: todo +created-todos sent pageSize=50 while
the backend silently returns empty for pageSize>20; now uses the shared pager
(pageSize=20).
Affected: contact/oa/wiki/drive/minutes/calendar/attendance/chat/report/smart
resolvers. Every fix ships a guard test that feeds the real backend response
shape (and, for minutes, both the taskUuid and minutesId item shapes) and
asserts the projection is non-empty with a usable id.
scripts/shortcut_real_result.py now compares the upper (projection) output
against the lower (raw backend) layer, and record_real_shortcut_run.py captures
the lower layer in memory (persisting only derived counts, never raw PII) so an
exit-0 empty projection over a non-empty backend is scored as
projection-data-loss instead of real-ok. The Python self-test runs in CI via
test/scripts.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Normalize message projections across read shortcuts, preserve mixed user JSON, expand forwarded records, mask ciphertext, and accept media-download message ID aliases while retaining the Cobra/Schema required contract.
Replace release-hidden terminology with a generated public shortcut catalog, remove include-hidden discovery, and keep real-test followups as an internal CR artifact.
