docs: add agent-readable document context workflow - #2121
Conversation
|
No actionable comments were generated in the recent review. 🎉 ℹ️ Recent review info⚙️ Run configurationConfiguration used: defaults Review profile: CHILL Plan: Pro Plus Run ID: 📒 Files selected for processing (2)
🚧 Files skipped from review as they are similar to previous changes (1)
📝 WalkthroughWalkthroughThe changes refine lark-doc’s document-context reading rules and update quick-decision guidance for deep-reading requests, including synced references, embedded Sheet/Base content, keyword targeting, metadata extraction, and request deduplication. ChangesLark document context workflow
Estimated code review effort: 1 (Trivial) | ~3 minutes Possibly related PRs
Suggested labels: Suggested reviewers: 🚥 Pre-merge checks | ✅ 5✅ Passed checks (5 passed)
✨ Finishing Touches🧪 Generate unit tests (beta)
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
There was a problem hiding this comment.
Actionable comments posted: 4
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Inline comments:
In `@skills/lark-doc/references/lark-doc-context.md`:
- Around line 60-61: Update the read deduplication guidance to key requests by
full request identity rather than token alone, including differing revisions,
scopes, ranges, detail levels, or fields. Only deduplicate equivalent
document/resource reads, while preserving the existing partial-failure behavior
that retains retrieved content and marks uncovered data.
- Around line 18-21: Update the document-reading workflow in the fetch-range
guidance to prioritize the keyword path when the user supplies a concrete search
term, using the keyword option from lark-doc-fetch.md before
outline/section/range. Preserve the existing simple-detail path for short
overall-document requests and the outline-based path for long or partial
requests without a specific keyword.
- Around line 45-53: The embedded-resource policy is inconsistent between the
context workflow and quick-decision guidance. In
skills/lark-doc/references/lark-doc-context.md lines 45-53, define whether
Sheet/Base tags require mandatory drill-down or only task-relevant conditional
processing; then update the quick-decision trigger in skills/lark-doc/SKILL.md
line 38 to use the same rule and remain consistent with the existing
embedded-resource guidance.
In `@skills/lark-doc/SKILL.md`:
- Line 38: Update the quick-decision rule in SKILL.md to include synced
references as a trigger for reading lark-doc-context.md and orchestrating the
existing read-only capabilities. Preserve the current conditions for comments,
images, whiteboards, and embedded Sheet/Base content, or replace them with
wording that explicitly covers all relevant structured references.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: defaults
Review profile: CHILL
Plan: Pro Plus
Run ID: e2bf36ae-4b13-4943-abea-9e9b59f225eb
📒 Files selected for processing (3)
skills/lark-doc/SKILL.mdskills/lark-doc/references/lark-doc-context.mdskills/lark-doc/references/lark-doc-fetch.md
Summary
Document how agents can turn one deep-read intent into a bounded composition of existing read-only commands, while preserving the docs/drive/sheets/base domain boundaries.
Changes
Test Plan
Related Issues
Summary by CodeRabbit