agent-skills
AGENTS.md2 starsChanged 17 days agoMITView on GitHub
This repository publishes reusable Agent Skills. Keep the repo easy for installers and agents to scan. - Put every public skill in skills/<skill-name>/. - Use lowercase kebab-case for skill folders. - Every skill must contain SKILL.md. - Optional supporting resources belong in scripts/, references/, assets/, or agents/. - Do not add README files inside individual skill directories; put public docs in the repo README and agent instructions in SKILL.md. - SKILL.md frontmatter must contain name and description. Extra fields are…
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 31,843 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 repository22 other files it gives its agents.
- act-within-scopeSkill
- build-filesSkill
- codingSkill
- context-managementSkill
- create-verification-skillSkill
- domain-modelingSkill
- encode-lessonsSkill
- experience-firstSkill
- eyes-onSkill
- fluxSkill
- frontend-designSkill
- gitSkill
- grill-with-docsSkill
- in-your-own-wordsSkill
- lint-rulesSkill
- maintain-verification-skillSkill
- scripted-verificationSkill
- security-auditSkill
- skill-evaluationSkill
- test-auditSkill
- test-behavior-not-implementationSkill
- writingSkill
The file
We split this AGENTS.md into the rules it gives an agent
11 rules. About 0 are common, 5 are distinctive, and 6 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈5
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
- Very specific: line 8
Put every public skill in skills/<skill-name>/.
- Distinctive: line 9
Use lowercase kebab-case for skill folders.
- Very specific: line 10
Every skill must contain SKILL.md.
- Very specific: line 13
Do not add README files inside individual skill directories.
- Very specific: line 23
Put all activation and routing language in description.
- Distinctive: line 24
Prefer concise, imperative instructions in the body.
- Distinctive: line 27
Add scripts only when they make repeated or fragile work more reliable.
- Very specific: line 28
When adding a new skill folder, run npm run link-skills so the local ~/.agents/skills/<skill-name> entry exists before validation.
- Distinctive: line 33
Run this before publishing changes:
- Distinctive: line 45
Always prepare release notes when cutting a release unless the user explicitly says to skip release notes.
- Very specific: line 47
If changelog or release-note evidence exists (CHANGELOG.md, docs/releases/, docs/changelog/, releases/, or .github/releases/), update it …
≈ “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
Agent Skills Repository Guide
This repository publishes reusable Agent Skills. Keep the repo easy for installers and agents to scan.
Structure
- Put every public skill in
skills/<skill-name>/. - Use lowercase kebab-case for skill folders.
- Every skill must contain
SKILL.md. - Optional supporting resources belong in
scripts/,references/,assets/, oragents/. - Do not add README files inside individual skill directories; put public docs
in the repo README and agent instructions in
SKILL.md.
Skill Rules
SKILL.mdfrontmatter must containnameanddescription. Extra fields are allowed when a consuming tool needs them. For manual-only skills usedisable-model-invocation: true(Cursor, Claude Code) and, for OpenCode,metadata.opencode/autoinvoke: "false"plusmetadata.opencode/slash: "true".- The
namevalue should match the folder name. - Put all activation and routing language in
description. - Prefer concise, imperative instructions in the body.
- Link directly from
SKILL.mdto supporting reference files so agents can load them only when relevant. - Add scripts only when they make repeated or fragile work more reliable.
- When adding a new skill folder, run
npm run link-skillsso the local~/.agents/skills/<skill-name>entry exists before validation.
Validation
Run this before publishing changes:
npm run validateThe validator checks skill folder names, frontmatter boundaries, required
fields, duplicate names, references to missing local markdown files, and local
entries in ~/.agents/skills for every skills/<skill-name>/ directory.
Releases
- Always prepare release notes when cutting a release unless the user explicitly says to skip release notes.
- If changelog or release-note evidence exists (
CHANGELOG.md,docs/releases/,docs/changelog/,releases/, or.github/releases/), update it before tagging. - In-repo changelog or release-note updates must be committed before tagging.
- Tag the commit that includes required release-note updates.
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.