| name | review |
| description | This skill should be used when the user asks to "review tldr notes", "triage tldr improvements", "go through tldr backlog", "check my tldr notes", "what tldr notes do I have", "/tldr:review", or wants a dedicated session to discuss, refine, and act on pending improvement notes captured via /tldr:note. Walks through each pending item for discussion and implementation. |
review
Guided triage session through pending TLDR improvement notes. Discuss, refine, and act on items captured via /tldr:note.
Purpose
Notes captured mid-workflow via /tldr:note are intentionally brief. Provide a dedicated session to:
- Review each pending note with full context
- Refine the agent's interpretation ("I said X but I meant Y")
- Decide on action: implement now, defer, or discard
- Implement changes to the TLDR skill when ready
Path Resolution
All data lives inside the TLDR plugin directory. This skill is at skills/review/SKILL.md — the plugin root is two directories up (the directory containing skills/, docs/, and .claude-plugin/). All file paths below are relative to the plugin root.
- Notes catalog:
docs/evaluation/notes.md
- TLDR skill file:
skills/tldr/SKILL.md
Review Workflow
Step 1: Load Pending Notes
Read docs/evaluation/notes.md and identify all entries with **Status:** pending.
If no pending notes exist, inform the user:
"No pending notes to review. Use /tldr:note [context] to capture improvement ideas."
Step 2: Present Summary
Show a brief overview of all pending notes:
Pending TLDR notes: N items
| # | Date | Category | Summary |
|---|------|----------|---------|
| 001 | 2026-02-15 | improvement | Numbered questions get flattened |
| 002 | 2026-02-15 | feature | Timeline format for debugging |
| ... | ... | ... | ... |
Start from the top, or pick a specific note number.
Step 3: Walk Through Each Note
For each pending note, present:
- The raw note — What the user originally captured
- The expanded interpretation — What the agent intuited
- The example — If one was attached via
--ex
- Related evaluation — If a sample was auto-created, reference its score and findings
Then ask the user:
"What would you like to do with this note?"
Options to present:
- Implement — Make the change to the TLDR skill now
- Refine — Correct the interpretation, add detail, then decide
- Defer — Keep as pending for a future session
- Discard — Remove (not relevant or already addressed)
Step 4: Handle Each Decision
Implement:
- Discuss the specific change needed
- Read the relevant TLDR skill file (
skills/tldr/SKILL.md or related files relative to the plugin root)
- Make the edit
- Update the note's status to
resolved with a resolution note
- Move to the next pending note
Refine:
- Let the user correct or expand the interpretation
- Update the note's "Expanded" field in
docs/evaluation/notes.md
- Re-present the updated note and ask for a decision again
Defer:
- Keep the note as
pending
- Optionally add user's comment about why it's deferred
- Move to the next pending note
Discard:
- Update the note's status to
discarded with a brief reason
- Move to the next pending note
Step 5: Session Summary
After processing all pending notes (or when the user wants to stop), provide a summary:
Review session complete:
- Implemented: N notes
- Deferred: N notes
- Discarded: N notes
- Remaining pending: N notes
Changes made:
- [List of specific edits to TLDR skill files]
Status Values
Notes progress through these statuses:
pending — Awaiting review (set by /tldr:note)resolved — Implemented or addressed (set during review)discarded — Not relevant or already fixed (set during review)
Interaction Style
- Conversational — Treat this as a discussion, not a checklist. Allow the user to explore ideas.
- One at a time — Present notes individually to avoid overwhelm.
- User drives pace — Let the user decide when to move on. Do not rush through items.
- Implementation ready — When the user says "implement", read the relevant files and make changes immediately. Do not just describe what to change.
