lark-note
Look up Lark meeting notes by note_id
This skill retrieves Lark meeting note details and transcripts when you already have the note_id. It avoids searching or guessing note IDs, ensuring you get accurate, verifiable results from official API calls.
Download the skill ZIP
Upload in Claude
Go to Settings → Capabilities → Skills → Upload skill
Toggle on and start using
Test it
Using "lark-note". Get details for note_id 7000000000000000000
Expected outcome:
- note_id: 7000000000000000000
- note_display_type: unified
- note_doc_token: doxcndocs123
- verbatim_doc_token: (empty or present depending on type)
Using "lark-note". Fetch the unified transcript for note_id 7000000000000000000
Expected outcome:
- File saved to ./notes/7000000000000000000/unified_transcript.md
- Contains full transcript with speaker attribution and timestamps
Security Audit
SafeThis skill contains only Markdown documentation files describing CLI command routing for Lark/Feishu meeting note lookups. The 112 static findings are all false positives caused by regex pattern matching on backtick characters in code formatting (e.g., `lark-cli note +detail`) and relative documentation links (e.g., `../lark-doc/SKILL.md`). No executable code, no network requests, no data exfiltration patterns, and no malicious intent detected.
Risk Factors
⚙️ External commands (4)
📁 Filesystem access (3)
Quality Score
What You Can Build
Look up meeting note details
Given a note_id, retrieve the note's display type, document token, and related metadata.
Fetch a unified transcript
Retrieve the full transcript of a unified meeting note, saved to a local markdown file.
Route from document to note
Extract vc-node-id from a Lark document's vc-transcribe-tab element and use it to query the note domain.
Try These Prompts
Look up the meeting note with note_id 1234567890.
Fetch the unified transcript for note_id abc123 and save it locally.
Read this Lark document, find the vc-transcribe-tab element, extract the vc-node-id, then look up the note details.
Check the note_display_type for note_id xyz789 and tell me whether it's unified, normal, or unknown.
Best Practices
- Always require an explicit note_id before invoking any command; never reverse-engineer from doc_token or title
- Check note_display_type first to determine the correct routing path for transcripts
- Use `--format json` for machine-readable output when integrating with other tools
Avoid
- Do not guess note_id from document title, body content, or backlink references
- Do not use `note +transcript` for normal-type notes; route to `docs +fetch` instead
- Do not add `--overwrite` without explicit user confirmation