* fix: remove sensitive send scope from reply and forward shortcuts
Remove mail:user_mailbox.message:send from the required scopes of
+reply, +reply-all, and +forward shortcuts. This scope is sensitive
and may not be granted, while these shortcuts default to saving
drafts and do not strictly require it.
* fix: validate send scope dynamically when --confirm-send is set
Add validateConfirmSendScope() to check mail:user_mailbox.message:send
in the Validate phase when --confirm-send is used, preventing the
"draft created but send failed" scenario. Add regression tests for
+reply, +reply-all, and +forward.
MarkRaw previously skipped both enrichPermissionError and
WriteErrorEnvelope, causing api command to exit 1 with no output
on API errors. Now MarkRaw only skips enrichPermissionError while
WriteErrorEnvelope always runs, ensuring stderr error envelope is
always written.
- Simplify apiRun MarkRaw logic (remove unnecessary IsJSONContentType check)
- Update existing tests to match new behavior
- Add 5 e2e tests covering api/service/shortcut error output
* ci: improve CI workflows and add golangci-lint config
- Add path filters to avoid unnecessary CI runs on non-Go changes
- Use go-version-file instead of hardcoded Go version
- Unify runners to ubuntu-latest
- Consolidate staticcheck/vet into golangci-lint with curated linter set
- Add go mod tidy check, govulncheck, and dependency license check
- Enable race detector in coverage, increase test timeout to 5m
- Add build verification step to tests workflow
- Add .codecov.yml with patch coverage target (60%)
- Add .golangci.yml (v2) with security and correctness linters
Change-Id: I409beb21cc1f1568ff47739c0a00f6214c10a0dd
* ci: replace Codecov upload with GitHub Job Summary coverage report
- Remove Codecov action dependency and CODECOV_TOKEN usage
- Generate coverage report using go tool cover and display in Job Summary
- Rename job from 'codecov' to 'coverage'
- Remove .codecov.yml from paths filter
Change-Id: Ib65dab6c4d7117c3300a9ea31eb1550537c72f88
* ci: trigger lint workflow
Change-Id: Ic1c492dd339f5460d2be2971ac65ea8f99e524eb
* ci: replace golangci-lint action with go run to avoid action whitelist restriction
Change-Id: I87274abf9780eb8b6350e98a27302ec5acc2a2e5
* ci: replace golangci-lint action with go run, keep incremental lint via --new-from-rev
Change-Id: I3d4a13cfd7b6c02e4098b04b8533a7248185c077
* ci: add fetch-depth 0 to lint checkout for incremental lint to work
Change-Id: I112279c5ec06dc0aa3aa7e01d564ea27fbd20533
* ci: disable errcheck linter due to high volume of existing violations
Change-Id: Iec57e8fbe42699f687d931d9dde2f879f2ae5b02
* ci: align golangci-lint config with GitHub CLI, make govulncheck non-blocking
- Add exptostd, gocheckcompilerdirectives, gochecksumtype, gomoddirectives linters
- Move gosec, staticcheck, errname, errorlint, misspell to TODO for later enablement
- Remove G104 exclusion (errcheck is disabled)
- Make govulncheck continue-on-error until Go version is upgraded
Change-Id: I330ece4f202229aee1e2f50790f6b22738704c05
* ci: fix go-licenses module path for v2
Change-Id: Ifd018ebe79cd18402171417b1b73313af2d23c6d
Combine the separate "Install CLI" and "Install AI Agent Skills" sections
into a single "Install" section for both human and AI Agent quick start
guides. Renumber the AI Agent steps accordingly.
Change-Id: I4ac10538d912a8889f52ea5a85e757d3e8bad21e
Add npm version badge to both README files for better package
discoverability. Reword the Quick Start tip to directly address
AI assistants instead of human users.
Change-Id: I9fb4252e4a7bde4ab6644c6ca6e63dc5d34b6f0c
Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
* docs: emphasize Skills installation as required step for AI Agents
Split the AI Agent quick start from a single code block into separate
labeled steps, and add a prominent note explaining that without Skills
the Agent cannot discover available commands or parameters.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
* docs: position Skills installation as the core step of the setup
Reword the Skills step to convey that it is the most important part of
the entire setup — the CLI is just a binary, Skills are what give the
Agent the knowledge to operate Lark.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
* docs: tone down Skills step wording, keep it concise
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
feng zhi hao