agentleFS
jara-r-kjara-r-k/claude-code-skills/CLAUDE.md

claude-code-skills

CLAUDE.md1 starsChanged 46 days agoMITView on GitHub

Root principles: See /Users/jararokahrknowles/projects/AGENTS.md §1–4 (Think Before Coding, Simplicity First, Surgical Changes, Goal-Driven Execution). They apply here — especially §2 (skills should be minimal, not speculative) and §4 (validate against compliance checklist before publishing). Public repository of reusable Claude Code skills, agents, and examples. Every SKILL.md requires YAML frontmatter: Body must be < 5,000 words. Include ## Error Handling and ## Examples sections. Include negative triggers (what the skill should NOT do). Agents use YAML frontmatter with name, model (Sonnet…

The file

We split this CLAUDE.md into the rules it gives an agent

3 rules. About 0 are common, 1 is distinctive, and 2 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈1

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 2

    Very specific

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

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

    Every SKILL.md requires YAML frontmatter:

    line 20
  2. Very specific:

    All skills must pass /skill-compliance-checklist before publishing

    line 43
  3. Distinctive:

    When compacting, preserve: list of modified skills with compliance status, active gotchas, and version numbers.

    line 80

≈ “Common” is an estimate: at least 1 in 5 of the 26,980 CLAUDE.md files we read have a rule of the same kind, not the same words. “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
CLAUDE.md

Claude Code Skills

Root principles: See /Users/jararokahrknowles/projects/AGENTS.md §1–4 (Think Before Coding, Simplicity First, Surgical Changes, Goal-Driven Execution). They apply here — especially §2 (skills should be minimal, not speculative) and §4 (validate against compliance checklist before publishing).

Public repository of reusable Claude Code skills, agents, and examples.

Structure

skills/             # SKILL.md files in kebab-case dirs
  {name}/SKILL.md   # Primary skill definition
  {name}/references/ # Supporting docs (optional)
agents/             # Reusable agent definitions (.md)
examples/           # Adaptation templates with SKILL.md + references/
.github/workflows/  # CI: validate-skills.yml

Skill Format

Every SKILL.md requires YAML frontmatter:

---
name: kebab-case-name        # required, kebab-case only
description: "< 1024 chars"  # required, no XML angle brackets
license: MIT
user-invocable: true|false
metadata:
  author: jara-r-k
  version: X.Y.Z
  based-on: "source reference"
---

Body must be < 5,000 words. Include ## Error Handling and ## Examples sections. Include negative triggers (what the skill should NOT do).

Agent Format

Agents use YAML frontmatter with name, model (Sonnet default), and tools list. Body describes methodology and output format.

Conventions

  • All skills must pass /skill-compliance-checklist before publishing
  • Kebab-case naming for all directories and skill names
  • Australian English spelling throughout
  • MIT licence for all public content
  • Version bump in frontmatter on every functional change
  • Git: conventional commits (feat:, fix:, docs:)

Development Workflow

# No build step — Markdown only
# Validation is CI-only via GitHub Actions
# Local check: run skill-compliance-checklist skill against target

Success Criteria Patterns

TaskDone when
New skillSKILL.md has valid frontmatter, < 5K words, includes error handling + examples + negative triggers, CI passes
Edit skillExisting frontmatter preserved, version bumped, CI passes, compliance re-validated
New agentFrontmatter has name/model/tools, methodology section present, CI passes
New exampleSKILL.md + references/ dir, customisation guide in references, CI passes

Gotchas & Failure Modes

  1. Description > 1024 chars — CI fails silently-ish. Check length before pushing.
  2. XML angle brackets in description — <tool> in description field breaks parsing. Use backticks in body, never in frontmatter description.
  3. Non-kebab-case names — CI rejects uppercase, underscores, spaces in name: field.
  4. Word count creep — Skills over 5,000 words fail CI. Split into skill + references/ if approaching limit.
  5. Missing error handling section — CI warns but doesn't fail. Include it anyway for compliance.
  6. Frontmatter delimiter — Must be exactly --- on line 1. No leading whitespace.
  7. Agent files skip word count — CI only checks frontmatter for agents, not body length. Self-enforce.
  8. Wiki raw/ safety — This project is part of the ~/projects/ LLM Wiki. Never write to ~/projects/raw/ — that layer is human-curated and immutable. See /Users/jararokahrknowles/projects/AGENTS.md for the full wiki schema.

Compact Instructions

When compacting, preserve: list of modified skills with compliance status, active gotchas, and version numbers.

Wiki

Project entity: [[entities/claude-code-skills]] in ~/projects/wiki/

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.