Agent skill
retro-create
Post-invocation retrospective. Tracks skill issues and applies fixes automatically.
Install this agent skill to your Project
npx add-skill https://github.com/majiayu000/claude-skill-registry/tree/main/skills/other/other/retro-create
SKILL.md
Purpose
After any skill completes, review the session for issues and improve the skill to prevent recurrence.
Trigger
Run automatically after every skill invocation completes (success or failure).
What to look for
- Failed commands — commands that errored and required retrying or workarounds
- User corrections — the user stepped in to redirect, correct, or override
- Wrong assumptions — the skill assumed something that was not true (versions, paths, flags)
- Missing steps — manual steps the user had to perform that the skill should have handled
- Outdated instructions — references to APIs, flags, or behaviors that have changed
Protocol
For each issue found:
- Log — append an entry to
<skill-dir>/retro.md(create the file if it does not exist) - Root-cause — identify which file (SKILL.md, reference, template) contains the gap
- Fix immediately — edit the skill file directly to prevent recurrence. This is the default action.
- Log only if unsafe — record the issue with
status: openonly when the fix would change skill behavior in ways that cannot be verified in the current session. Require aresolution-hinton every open entry.
After logging any status: open entry, append:
Open issues remain. Run /retro-resolve to process them.
retro.md entry format
## YYYY-MM-DD — <brief title>
- **type**: failed-command | user-correction | wrong-assumption | missing-step | outdated-instruction
- **skill**: <skill-name>
- **context**: <what happened>
- **root-cause**: <why it happened>
- **fix**: <what was changed> | `status: open` if not yet fixed
- **resolution-hint**: <what the fix should look like> (required for open entries)
- **files-changed**: <list of files edited, or "none">
Constraints
- Do not modify skill behavior or semantics — only fix gaps, errors, and omissions
- Keep entries concise (3-5 lines per field max)
- Do not duplicate fixes already present in skill files
- If a fix would change skill behavior significantly, log as
status: openand note the concern - Prefer fixing over logging. If you can describe the fix in
resolution-hint, you can probably just apply it. - Never log as
status: opensimply because the session is long or context is low.
Recommended Agent Skills
Expand your agent's capabilities with these related and highly-rated skills.
agent-ops-spec
Manage specification documents in .agent/specs/. Use when user provides requirements, acceptance criteria, or feature descriptions that need to be tracked and validated against implementation.
agent-ops-state
Maintain .agent state files. Use at session start, after meaningful steps, and before concluding: read/update constitution/memory/focus/issues/baseline consistently.
agent-ops-spec
Manage specification documents in .agent/specs/. Use when user provides requirements, acceptance criteria, or feature descriptions that need to be tracked and validated against implementation.
agent-ops-testing
Test strategy, execution, and coverage analysis. Use when designing tests, running test suites, or analyzing test results beyond baseline checks.
agent-ops-testing
Test strategy, execution, and coverage analysis. Use when designing tests, running test suites, or analyzing test results beyond baseline checks.
agent-ops-state
Maintain .agent state files. Use at session start, after meaningful steps, and before concluding: read/update constitution/memory/focus/issues/baseline consistently.
Didn't find tool you were looking for?