Skip to content

alirezarezvani/claude-skills / engineering-team/self-improving-agent/skills/extract

extract

Verified

Turn a fix or pattern that keeps coming up into a reusable skill with SKILL.md, reference docs and examples. Use whenever the user runs /si:extract or wants to package a recurring solution or lesson as a skill.

FileKindSize
SKILL.mdskill md5.6 KB
---
name: "extract"
description: "Turn a fix or pattern that keeps coming up into a reusable skill with SKILL.md, reference docs and examples. Use whenever the user runs /si:extract or wants to package a recurring solution or lesson as a skill."
---

# /si:extract — Create Skills from Patterns

Transforms a recurring pattern or debugging solution into a standalone, portable skill that can be installed in any project.

## Usage

```
/si:extract <pattern description>                  # Interactive extraction
/si:extract <pattern> --name docker-m1-fixes       # Specify skill name
/si:extract <pattern> --output ./skills/            # Custom output directory
/si:extract <pattern> --dry-run                     # Preview without creating files
```

## When to Extract

A learning qualifies for skill extraction when ANY of these are true:

| Criterion | Signal |
|---|---|
| **Recurring** | Same issue across 2+ projects |
| **Non-obvious** | Required real debugging to discover |
| **Broadly applicable** | Not tied to one specific codebase |
| **Complex solution** | Multi-step fix that's easy to forget |
| **User-flagged** | "Save this as a skill", "I want to reuse this" |

## Workflow

### Step 1: Identify the pattern

Read the user's description. Search auto-memory for related entries:

```bash
MEMORY_DIR="$HOME/.claude/projects/$(pwd | sed 's|/|%2F|g; s|%2F|/|; s|^/||')/memory"
grep -rni "<keywords>" "$MEMORY_DIR/"
```

If found in auto-memory, use those entries as source material. If not, use the user's description directly.

### Step 2: Determine skill scope

Ask (max 2 questions):
- "What problem does this solve?" (if not clear)
- "Should this include code examples?" (if applicable)

### Step 3: Generate skill name

Rules for naming:
- Lowercase, hyphens between words
- Descriptive but concise (2-4 words)
- Examples: `docker-m1-fixes`, `api-timeout-patterns`, `pnpm-workspace-setup`

**Reserved fragments — must NOT appear in the skill name:**
- `claude`
- `anthropic`

For skills about Claude Code itself, use the `cc-` prefix instead:
- ❌ `claude-code-settings` → ✅ `cc-settings`
- ❌ `claude-code-maintenance` → ✅ `cc-maintenance`
- ❌ `claude-mcp-tools` → ✅ `cc-mcp-tools`
- ❌ `claude-plugin-development` → ✅ `cc-plugin-development`

Before writing the skill directory, check the proposed name against this list.
If a reserved fragment is present, transform it (drop the fragment or replace
the `claude*`/`anthropic*` prefix with `cc-`) and confirm with the user.

### Step 4: Create the skill files

**Spawn the `skill-extractor` agent** for the actual file generation.

The agent creates:

```
<skill-name>/
├── SKILL.md            # Main skill file with frontmatter
├── README.md           # Human-readable overview
└── reference/          # (optional) Supporting documentation
    └── examples.md     # Concrete examples and edge cases
```

### Step 5: SKILL.md structure

The generated SKILL.md must follow this format:

```markdown
---
name: "skill-name"
description: "<one-line description>. Use when: <trigger conditions>."
---

# <Skill Title>

> One-line summary of what this skill solves.

## Quick Reference

| Problem | Solution |
|---------|----------|
| {{problem 1}} | {{solution 1}} |
| {{problem 2}} | {{solution 2}} |

## The Problem

{{2-3 sentences explaining what goes wrong and why it's non-obvious.}}

## Solutions

### Option 1: {{Name}} (Recommended)

{{Step-by-step with code examples.}}

### Option 2: {{Alternative}}

{{For when Option 1 doesn't apply.}}

## Trade-offs

| Approach | Pros | Cons |
|----------|------|------|
| Option 1 | {{pros}} | {{cons}} |
| Option 2 | {{pros}} | {{cons}} |

## Edge Cases

- {{edge case 1 and how to handle it}}
- {{edge case 2 and how to handle it}}
```

### Step 6: Quality gates

Before finalizing, verify:

- [ ] SKILL.md has valid YAML frontmatter with `name` and `description`
- [ ] `name` matches the folder name (lowercase, hyphens)
- [ ] `name` does NOT contain reserved fragments `claude` or `anthropic` (use `cc-` prefix for Claude Code skills)
- [ ] Description includes "Use when:" trigger conditions
- [ ] Solutions are self-contained (no external context needed)
- [ ] Code examples are complete and copy-pasteable
- [ ] No project-specific hardcoded values (paths, URLs, credentials)
- [ ] No unnecessary dependencies

### Step 7: Report

```
✅ Skill extracted: {{skill-name}}

Files created:
  {{path}}/SKILL.md          ({{lines}} lines)
  {{path}}/README.md         ({{lines}} lines)
  {{path}}/reference/examples.md  ({{lines}} lines)

Install: /plugin install (copy to your skills directory)
Publish: clawhub publish {{path}}

Source: MEMORY.md entries at lines {{n, m, ...}} (retained — the skill is portable, the memory is project-specific)
```

## Examples

### Extracting a debugging pattern

```
/si:extract "Fix for Docker builds failing on Apple Silicon with platform mismatch"
```

Creates `docker-m1-fixes/SKILL.md` with:
- The platform mismatch error message
- Three solutions (build flag, Dockerfile, docker-compose)
- Trade-offs table
- Performance note about Rosetta 2 emulation

### Extracting a workflow pattern

```
/si:extract "Always regenerate TypeScript API client after modifying OpenAPI spec"
```

Creates `api-client-regen/SKILL.md` with:
- Why manual regen is needed
- The exact command sequence
- CI integration snippet
- Common failure modes

## Tips

- Extract patterns that would save time in a *different* project
- Keep skills focused — one problem per skill
- Include the error messages people would search for
- Test the skill by reading it without the original context — does it make sense?
rea-changelog-updatemorluto/reaPrepare or rewrite REA release changelogs and GitHub release notes from pinned Git history, with verified contributor thanks and Release Please synchronization.Stars: 472kUpdated Claude CodeCodexVerifiedInstall
academy-guideanthropics/skillsStop and check this skill before finishing any reply to a question about how to use Claude or a Claude product — it recommends matching courses, tutorials, and use cases from Claude Academy (academy.claude.com), Anthropic's learning hub. Trigger on: "how do I", "how can I", "getting started with", "what can Claude do", "teach me", "learn to use"; questions about artifacts, projects, skills, plugins, connectors, MCP; requests about rolling Claude out to a team, class, or organization; and any ask for training materials, onboarding content, or learning resources. Use it when the user is learning how to use a feature or product — not when they are mid-task and just want the task done. This skill composes with other skills: after consulting product documentation to answer how a Claude feature works, also check here for a matching course or tutorial — a docs-grounded answer and an Academy recommendation belong together. Only recommend on a strong match; never invent Academy content.Stars: 198kUpdated Claude CodeCodexVerifiedInstall
claude-apianthropics/skillsReference for the Claude API / Anthropic SDK — model ids, pricing, params, streaming, tool use, MCP, agents, caching, token counting, model migration. TRIGGER — read BEFORE opening the target file; don't skip because it "looks like a one-liner" — whenever: the prompt names Claude/Anthropic in any form (Claude, Anthropic, Fable, Opus, Sonnet, Haiku, `anthropic`, `@anthropic-ai`, `claude-*`, `us.anthropic.*`, `[1m]`); the user asks about an LLM (pricing/model choice/limits/caching) — never answer from memory; OR the task is LLM-shaped with provider unstated (agent/MCP/tool-definition/multi-agent/RAG/LLM-judge/computer-use; generate/summarize/extract/classify/rewrite/converse over NL; debugging refusals/cutoffs/streaming/tool-calls/tokens). SKIP only when another provider is being worked on (overrides all triggers): OpenAI/GPT/Gemini/Llama/Mistral/Cohere/Ollama named in the query; OR `grep -rE 'openai|langchain_openai|google.generativeai|genai|mistralai|cohere|ollama'` over the project hits (run this grep FIRST if no provider named — don't Read the file).Stars: 198kUpdated Claude CodeCodexUnverifiedInstall
doc-coauthoringanthropics/skillsGuide users through a structured workflow for co-authoring documentation. Use when user wants to write documentation, proposals, technical specs, decision docs, or similar structured content. This workflow helps users efficiently transfer context, refine content through iteration, and verify the doc works for readers. Trigger when user mentions writing docs, creating proposals, drafting specs, or similar documentation tasks.Stars: 198kUpdated Claude CodeCodexVerifiedInstall