skills
AGENTS.md4 starsChanged 3 months agoMITView on GitHub
Public collection of reusable Codex skills and slash-command workflows. Each top-level skill is an independently installable folder, not application source.
Where files are headingNew AGENTS.md files have more rules about keeping secrets out of the code, and fewer about code style.
Of the AGENTS.md files we read, the 31,774 first seen in the last six months, against the older ones: the kind of rule that gained most share, and the kind that lost most. About files like this in general, not this one, and an estimate, since a kind is a pattern in the words.
In this repository21 other files it gives its agents.
- 1Skill
- 2Skill
- 3Skill
- 4Skill
- 5Skill
- ae-extendscriptSkill
- asapSkill
- backbeforeSkill
- componentsSkill
- githubSkill
- huhSkill
- instantSkill
- motion-animateSkill
- motion-assetsSkill
- motion-designSkill
- motion-pipelineSkill
- motion-planSkill
- quick-contextSkill
- shader-uiSkill
- thumbnailSkill
- ui-intergritySkill
The file
We split this AGENTS.md into the rules it gives an agent
9 rules. About 1 is common, 5 are distinctive, and 3 are very specific to this project.
≈1
Common · estimate
Many files like this have a rule of this kind.
≈5
Distinctive · estimate
This author's own choices. Worth a look.
3
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- could be automatic: a setting could enforce it
- Distinctive: line 10
No repository-wide build or test command exists.
- Very specific: line 23
Keep each skill narrowly scoped with accurate name and description frontmatter.
- Distinctive: line 24
Read a changed skill completely before editing its supporting references or scripts.
- Very specific: line 25
Keep SKILL.md concise; move optional depth into references/ and route to it explicitly.
- Distinctive: line 26
Use relative links within a skill package so installed copies remain portable.
- Distinctive: line 27
Keep destructive/external actions explicit and guarded.
- Very specific: line 28
Update matching agents/openai.yaml metadata when skill identity or invocation changes.
- Common: line 29
Do not add generated caches, local credentials, or machine-specific paths.
- Distinctive: line 35
Update README.md when adding, removing, or renaming a public skill.
≈ “Common” is an estimate: at least 1 in 5 of the 40,246 AGENTS.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
AGENTS.md
Project
Public collection of reusable Codex skills and slash-command workflows. Each top-level skill is an independently installable folder, not application source.
Commands
- No repository-wide build or test command exists.
- Validate changed helper scripts directly;
github/scripts/contains the current Python helpers.
Structure
<skill>/SKILL.md— required skill instructions and frontmatter.<skill>/agents/openai.yaml— OpenAI-facing metadata when supplied.github/references/— detailed GitHub workflow, safety, research, and template guidance.github/scripts/— repository audit and remote-guard helpers.README.md— public inventory; keep it synchronized with top-level skill folders.
Rules
- Keep each skill narrowly scoped with accurate
nameanddescriptionfrontmatter. - Read a changed skill completely before editing its supporting references or scripts.
- Keep
SKILL.mdconcise; move optional depth intoreferences/and route to it explicitly. - Use relative links within a skill package so installed copies remain portable.
- Keep destructive/external actions explicit and guarded.
- Update matching
agents/openai.yamlmetadata when skill identity or invocation changes. - Do not add generated caches, local credentials, or machine-specific paths.
Before finishing
- Re-read the changed skill as an agent would, following every referenced relative path.
- Confirm frontmatter parses and all referenced files exist.
- Update
README.mdwhen adding, removing, or renaming a public skill.
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.