skills
CLAUDE.md2 starsChanged 10 months agoMITView on GitHub
The description field in SKILL.md must include two parts: 1. What it does: Brief description of the skill's functionality 2. When to use: Tell the agent when to trigger this skill Example format: Bad example (missing trigger condition): To reduce token consumption, use progressive disclosure when a skill has many scripts or complex documentation: Example INDEX.md structure:
In this repository3 other files it gives its agents.
- mac-notifierSkill
- mermaid-to-imgSkill
- text-to-audioSkill
The file
We split this CLAUDE.md into the rules it gives an agent
8 rules. About 0 are common, 2 are distinctive, and 6 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈2
Distinctive · estimate
This author's own choices. Worth a look.
6
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 5
Commit messages must be in English
- Very specific: line 10
Use lowercase with hyphens (kebab-case), e.g., mac-notifier, mermaid-to-img
- Distinctive: line 12
When renaming a skill, update all three locations:
- Very specific: line 19
When adding a new skill, update both README.md (Chinese) and README_EN.md (English)
- Very specific: line 20
Add the new skill to the "Available Skills" table in both files
- Very specific: line 42
For skills with 3+ scripts: Create a scripts/INDEX.md file that briefly describes each script's purpose
- Very specific: line 43
Main SKILL.md: Only include the most common usage and reference INDEX.md for details
- Very specific: line 44
Detailed docs: Put comprehensive documentation in separate files (e.g., docs/ folder)
≈ “Common” is an estimate: at least 1 in 5 of the 27,641 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.
Files
Project Guidelines
Commit Rules
- Commit messages must be in English
Skill Naming Convention
- The skill directory name, SKILL.md
namefield, and plugin name must be identical - Use lowercase with hyphens (kebab-case), e.g.,
mac-notifier,mermaid-to-img - This ensures consistency between
/skill-namecommands and plugin management - When renaming a skill, update all three locations:
- Directory name
- SKILL.md
namefield .claude-plugin/marketplace.jsonpluginnameandskillspath
README Updates
- When adding a new skill, update both
README.md(Chinese) andREADME_EN.md(English) - Add the new skill to the "Available Skills" table in both files
Skill Description Rules
The description field in SKILL.md must include two parts:
- What it does: Brief description of the skill's functionality
- When to use: Tell the agent when to trigger this skill
Example format:
description: Convert Mermaid diagrams to images. Use when user wants to export/save/convert Mermaid code to PNG/JPG/SVG/PDF format.Bad example (missing trigger condition):
description: Convert Mermaid diagrams to images.Progressive Disclosure
To reduce token consumption, use progressive disclosure when a skill has many scripts or complex documentation:
- For skills with 3+ scripts: Create a
scripts/INDEX.mdfile that briefly describes each script's purpose - Main SKILL.md: Only include the most common usage and reference INDEX.md for details
- Detailed docs: Put comprehensive documentation in separate files (e.g.,
docs/folder)
Example INDEX.md structure:
# Scripts Index
| Script | Purpose |
|--------|---------|
| main.py | Primary entry point for basic usage |
| install.sh | Installation and setup |
| utils.py | Internal utilities (not for direct use) |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
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.

