agentleFS
theyoungastronautstheyoungastronauts/polaris/skills/writing-skills/SKILL.md

writing-skills

Skill4 starsChanged 4 months agoMITView on GitHub

How to author and structure Claude Code skills: SKILL.md frontmatter, writing trigger-focused descriptions, file organization, and naming. Apply when creating or editing a skill (a SKILL.md file).

191 other projects use a version of this, 2 of them edited.

Adopted by

Most recently taken up by

Used with

Projects with it also use chinese-code-review, chinese-commit-conventions and 10 more.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • Leodorareluctant259Skill

    chinese-code-review

    中文代码审查规范——在保持专业严谨的同时,用符合国内团队文化的方式给出有效反馈

    Also in 8 projects that use this

  • Leodorareluctant259Skill

    chinese-commit-conventions

    中文 Git 提交规范 — 适配国内团队的 commit message 规范和 changelog 自动化

    Also in 8 projects that use this

  • obraSkill

    diagnosing-superpowers

    Use when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness.

    Also in 20 projects that use this

  • kntismSkill

    condition-based-waiting

    Use when tests have race conditions, timing dependencies, or inconsistent pass/fail behavior - replaces arbitrary timeouts with condition polling to wait for actual state changes, eliminating flaky tests from timing guesses

    Also in 3 projects that use this

  • aaddrickSkill

    writing-agents

    Use when creating new agents, editing existing agents, or defining specialized subagent roles for the Task tool

    Also in 3 projects that use this

  • frankxaiSkill

    defense-in-depth

    Use when invalid data causes failures deep in execution, requiring validation at multiple system layers - validates at every layer data passes through to make bugs structurally impossible

    Also in 3 projects that use this

  • +3 more from Leodorareluctant259

The file

We split this skill into the rules it gives an agent

12 rules. 2 are common, 9 are distinctive, and 1 is very specific to this project.

  • 2

    Common

    Most of the 181 edited versions we read keep these.

  • 9

    Distinctive

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Every skill is a skills/<skill-name>/SKILL.md file that opens with YAML frontmatter, then markdown:

    line 44
  2. Distinctive:

    Description: Start with "Use when..." to focus on triggering conditions.

    line 77
  3. Distinctive:

    Keywords: Use words Claude would search for — error messages, symptoms, synonyms, tool names.

    line 87
  4. Distinctive:

    Every token in frequently-loaded skills counts.

    line 93
  5. Distinctive:

    Use cross-references instead of repeating content

    line 96
  6. Common:

    Don't repeat what's in cross-referenced skills

    line 98
  7. Distinctive:

    Keep content inline unless it exceeds ~100 lines of reference material or is a reusable script/utility.

    line 114
  8. Distinctive:

    Before deploying a skill, validate it works:

    line 118
  9. Distinctive:

    Baseline: Run a scenario WITHOUT the skill — document what the agent does wrong

    line 120
  10. Distinctive:

    With skill: Run the same scenario WITH the skill — verify the agent now complies

    line 121
  11. Distinctive:

    Close loopholes: If the agent finds workarounds, add explicit counters

    line 123
  12. Common:

    Untested skills: If you didn't watch an agent fail without the skill, you don't know if it teaches the right thing

    line 137

“Common” is measured: kept by at least 6 in 10 of the 181 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Writing Skills (Meta-Skill)

Adapted from obra/superpowers

Overview

A guide for creating effective, discoverable, and well-tested skills. Use this when authoring new skills for this repo.

What is a Skill?

A skill is a reference guide for proven techniques, patterns, or tools. Skills help AI agents find and apply effective approaches.

Skills are: Reusable techniques, patterns, tools, reference guides Skills are NOT: Narratives about how you solved a problem once

When to Create a Skill

Create when:

  • Technique wasn't intuitively obvious
  • You'd reference this again across projects
  • Pattern applies broadly (not project-specific)
  • Others would benefit

Don't create for:

  • One-off solutions
  • Standard practices well-documented elsewhere
  • Project-specific conventions (put in CLAUDE.md instead)
  • Mechanical constraints enforceable with linting/validation

Skill Types

  • Technique: Concrete method with steps (condition-based-waiting, root-cause-tracing)
  • Pattern: Way of thinking about problems (flatten-with-flags, test-invariants)
  • Reference: API docs, syntax guides, tool documentation

SKILL.md Structure

Every skill is a skills/<skill-name>/SKILL.md file that opens with YAML frontmatter, then markdown:

---
name: skill-name
description: One line — lead with the trigger use case (this drives auto-invocation).
# disable-model-invocation: true   # command-only: won't auto-trigger; run as /skill-name
# paths: ["**/*.py"]               # auto-trigger only when working on matching files
---

# Skill Name

## Overview
What is this? Core principle in 1-2 sentences.

## When to Use
Bullet list with SYMPTOMS and use cases.
When NOT to use.

## Core Pattern
Before/after comparison or step-by-step process.

## Quick Reference
Table or bullets for scanning common operations.

## Common Mistakes
What goes wrong + fixes.

Discovery Optimization

Future Claude instances need to FIND your skill. Optimize for this:

Description: Start with "Use when..." to focus on triggering conditions. Describe the problem, not the solution. NEVER summarize the skill's workflow in the description — Claude may follow the description shortcut instead of reading the full skill.

# BAD: Summarizes workflow
description: Use when executing plans - dispatches subagent per task with code review

# GOOD: Just triggering conditions
description: Use when implementing any feature or bugfix, before writing implementation code

Keywords: Use words Claude would search for — error messages, symptoms, synonyms, tool names.

Naming: Kebab-case directory names. For a command-only skill the directory name is the slash command (scaffold, verify-django, nextjs-bootstrap) — name it exactly what you'd type. For an auto-triggering skill use a clear noun/verb phrase that matches its trigger (django-patterns, work-discipline, plan-and-scope). Match the repo's existing names; don't force gerunds.

Token Efficiency

Every token in frequently-loaded skills counts.

  • Move details to supporting files, reference with cross-links
  • Use cross-references instead of repeating content
  • Compress examples — one excellent example beats many mediocre ones
  • Don't repeat what's in cross-referenced skills
  • Target: <200 words for frequently-loaded skills, <500 words for others

Code Examples

One excellent example beats many mediocre ones. Choose the most relevant language for the pattern. Make it complete, runnable, and well-commented. Don't implement in 5+ languages.

File Organization

skills/
  skill-name/
    SKILL.md              # Main reference (required)
    supporting-file.*     # Only if needed (heavy reference, reusable tools)

Keep content inline unless it exceeds ~100 lines of reference material or is a reusable script/utility.

Testing Skills

Before deploying a skill, validate it works:

  1. Baseline: Run a scenario WITHOUT the skill — document what the agent does wrong
  2. With skill: Run the same scenario WITH the skill — verify the agent now complies
  3. Edge cases: Try pressure scenarios where the agent might rationalize skipping the skill
  4. Close loopholes: If the agent finds workarounds, add explicit counters

Different skill types need different test approaches:

  • Discipline skills (rules/requirements): Test with pressure scenarios
  • Technique skills (how-to): Test with application scenarios and edge cases
  • Pattern skills (mental models): Test recognition and counter-examples
  • Reference skills (docs): Test retrieval and correct application

Anti-Patterns

  • Narrative style: "In session 2025-10-03, we found..." — too specific, not reusable
  • Multi-language dilution: examples in 5 languages — mediocre quality, maintenance burden
  • Code in flowcharts: Can't copy-paste, hard to read
  • Generic labels: "step1", "helper2" — labels should have semantic meaning
  • Untested skills: If you didn't watch an agent fail without the skill, you don't know if it teaches the right thing

Checklist for New Skills

  • Name uses letters, numbers, hyphens only
  • Clear overview with core principle in 1-2 sentences
  • "When to Use" section with specific triggers/symptoms
  • Keywords throughout for search discovery
  • One excellent code example (if applicable)
  • Quick reference table for scanning
  • Common mistakes section
  • Tested against baseline behavior
  • No narrative storytelling
  • Supporting files only for heavy reference or reusable tools

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.