skills
AGENTS.md0 starsChanged 6 days agoMITView on GitHub
This repo publishes agent skills for the Skills CLI (npx skills). It contains nothing else — no application code, no client material. - Flat layout. skills/<name>/SKILL.md. No category directories. - Directory name == name: in frontmatter. The CLI installs by name, so a mismatch breaks add and update. - Names are global. Installs land flat in ~/.agents/skills/<name>/ next to everyone else's skills. Pick a name unlikely to collide; avoid bare words like research, review, plan. - Harness built-in skills do…
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 32,671 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 repository9 other files it gives its agents.
- code-reviewSkill
- find-the-coreSkill
- glen-modeSkill
- north-starSkill
- show-meSkill
- start-with-whySkill
- ste-reportSkill
- wait-whatSkill
- wet-or-drySkill
The file
We split this AGENTS.md into the rules it gives an agent
4 rules. About 0 are common, 0 are distinctive, and 4 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈0
Distinctive · estimate
This author's own choices. Worth a look.
4
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 24
Relative paths only: Write references/foo.md, never an absolute path — the install location differs from the repo location.
- Very specific: line 26
description:is the trigger: Write it as the conditions under which the skill should fire, in the words a user would actually use. - Very specific: line 51
Add a row to the table in README.md — scripts/list-skills.sh --md prints it.
- Very specific: line 53
Run scripts/link-skills.sh to symlink it into ~/.claude/skills/ and ~/.agents/skills/, so edits here are live in both at once.
≈ “Common” is an estimate: at least 1 in 5 of the 41,581 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
Conventions for this repo
This repo publishes agent skills for the Skills CLI (npx skills). It contains
nothing else — no application code, no client material.
Rules
- Flat layout.
skills/<name>/SKILL.md. No category directories. - Directory name ==
name:in frontmatter. The CLI installs by name, so a mismatch breaksaddandupdate. - Names are global. Installs land flat in
~/.agents/skills/<name>/next to everyone else's skills. Pick a name unlikely to collide; avoid bare words likeresearch,review,plan. - Harness built-in skills do not count. Skills that ship inside a harness
(Claude Code, Codex, OpenCode, or any other) are not used, so choose names
and
description:triggers here as if they did not exist — the same name as a built-in is fine. - Only
SKILL.mdat a skill's root. Everything else goes inreferences/,scripts/,assets/, oragents/(harness metadata, such as Codex'sagents/openai.yaml). - Keep
SKILL.mdshort. It loads into context on every trigger. Past ~200 lines, move detail intoreferences/and link to it. Never duplicate a reference's content inSKILL.md. - Relative paths only. Write
references/foo.md, never an absolute path — the install location differs from the repo location. description:is the trigger. Write it as the conditions under which the skill should fire, in the words a user would actually use. It is the only part of the skill the agent sees before deciding to load it.- Use the glossary. Skills use the terms in
GLOSSARY.md. When a skill introduces or sharpens a term, update the glossary in the same change. - No confidential content. This repo is public. No client names, no HTGP or Ikoni internal material, no credentials, no personal data.
Frontmatter
Required:
---
name: my-skill
description: Use when <situation>. Covers <scope>. Not for <exclusion>.
---Optional, when they earn their place: argument-hint, user-invocable,
disable-model-invocation, allowed-tools, license, version.
Adding a skill
mkdir -p skills/<name>and writeSKILL.md.- Add a row to the table in
README.md—scripts/list-skills.sh --mdprints it. - Run
scripts/link-skills.shto symlink it into~/.claude/skills/and~/.agents/skills/, so edits here are live in both at once.
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.

