From taco
Retrospective analysis: extract learnings, update knowledge base, and refine rules
How this skill is triggered — by the user, by Claude, or both
Slash command
/taco:learnThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
```
/learn <feature-name>
Prerequisite: /review-code <feature-name> completed
Read the complete pipeline history:
.dev/specs/{feature-name}/idea.md.dev/specs/{feature-name}/spec.md.dev/specs/{feature-name}/plan.md.dev/specs/{feature-name}/reviews/plan-review.md.dev/specs/{feature-name}/reviews/code-review.mdAnalyze the pipeline execution:
What went well?
What didn't go well?
What to change?
Generate docs/learnings/{YYYY-MM-DD}-{feature-name}.md:
---
title: {Feature Name} — Learnings
tags: [{relevant tags}]
date: {YYYY-MM-DD}
feature: {feature-name}
---
## Summary
{1-2 sentence summary}
## What Went Well
{bullets}
## What Didn't Go Well
{bullets}
## Key Learnings
{numbered insights}
## Process Improvements
{suggestions for pipeline/template changes}
Run scripts/update-index.sh to add entry to docs/learnings/index.md.
If learnings suggest changes to:
Present proposed changes to user via AskUserQuestion. Only apply changes with explicit approval.
docs/learnings/, CLAUDE.md, and templates/.dev/specs/ retrospectivelynpx claudepluginhub mildsalmon/taco-harness --plugin tacoCaptures lessons learned from implementation, production, QA, and release so the project improves over time. Use after milestones, repeated failures, or postmortems.
Reviews completed work to extract learnings, validate shipping, and promote insights. Activates after tasks, PR arcs, or sessions finish, or after 5+ PRs.
Extracts reusable rules from developer corrections during a coding session and persists them to CLAUDE.md or memory for future sessions.