extract-approach

Extracts reasoning from solved problems into permanent learnings notes in docs/learnings/.

1|Updated May 23, 2026
One-click install
npx skills add https://github.com/palfish-t-i-u/palfish-t-i-u-h-th-ng-ver-2 --skill extract-approach-palfish-t-i-u
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: extract-approach
Source: https://github.com/palfish-t-i-u/palfish-t-i-u-h-th-ng-ver-2/tree/main/.claude/skills/extract-approach
Command: npx skills add https://github.com/palfish-t-i-u/palfish-t-i-u-h-th-ng-ver-2 --skill extract-approach-palfish-t-i-u

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? After solving a non-trivial bug, design decision, or incident, the reasoning behind the fix disappears when the session ends. This Skill captures the trap, insight, and rule behind a solution into a version-controlled Markdown note so future sessions and engineers skip the same discovery process. ## Core Features & Use Cases - Structured learnings notes: Writes notes to docs/learnings/ with five mandatory sections — Problem, Trap, Insight, Rule, and a runnable Verify command. - Index maintenance: Updates docs/learnings/README.md with a one-line index entry for each new note. - Quality gates: Enforces one insight per note, requires concrete file paths that exist, and rejects trivial problems that lack a real insight. - Use Case: After diagnosing why 10 of 14 FastAPI routes returned 404 due to an indentation shift in a route registration function, run this Skill to record the trap, the insight about silent include_router behavior, and a grep-based verify command. ## Quick Start Ask the AI to extract the approach from the bug fix just completed and save it as a learnings note in docs/learnings/.

Frequently Asked Questions about extract-approach

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I document lessons learned from a bug fix?

Run this Skill after completing a non-trivial fix. It writes a Markdown note to docs/learnings/ with five sections — Problem, Trap, Insight, Rule, and Verify — and adds a one-line entry to the docs/learnings/README.md index.

What should a good engineering learnings note contain?

A good note names concrete files and functions, states the obvious-but-wrong approach explicitly, captures the non-obvious insight, defines a checkable rule, and includes a runnable verify command such as a grep or curl that confirms the note is still valid.

When should I not write a learnings note?

Skip extraction for routine CRUD work, typo fixes, simple UI tweaks, or problems where the solution is obvious from the error message. If you cannot identify a genuine insight, the problem was too trivial to document.

Where are learnings notes stored and indexed?

Notes live in the version-controlled docs/learnings/ directory, readable by any engineer, model, or CI job. Each note gets a one-line entry under the Index heading in docs/learnings/README.md, not in the .claude memory system.

Why do stale learnings notes cause problems?

A note referencing a refactored file or moved line number misleads future readers. The Skill mitigates this by requiring every related file path to exist at write time and a runnable Verify command that detects when the note becomes outdated.