whanoftheworldwhanoftheworld/graphic-skills/AGENTS.md

graphic-skills

AGENTS.md0 starsChanged 34 days agoMITView on GitHub

Skills live flat in skills/<name>/SKILL.md, so every installer (Codex plugins, the skills CLI, copying by hand) finds them. Every skill is either user-invoked or model-invoked: Keep the two harness settings in sync; node scripts/check.mjs fails when they drift. A user-invoked skill can load a model-invoked one ("Load the x skill: the Skill tool, or read its SKILL.md") but can never fire another user-invoked skill: when a step needs one, tell the human to type it.

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 31,439 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 repository15 other files it gives its agents.
- ask-designerSkill
- brand-contextSkill
- brand-imageSkill
- brand-intakeSkill
- brand-systemSkill
- color-systemSkill
- design-briefSkill
- design-craftSkill
- design-critiqueSkill
- design-grillingSkill
- design-routesSkill
- grill-brandSkill
- logo-labSkill
- make-graphicSkill
- type-systemSkill

The file

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

3 rules. About 0 are common, 2 are distinctive, and 1 is 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.
1
Very specific
Tied to their project's names, files or tools. Probably skip.

Common
Distinctive
Very specific

Distinctive:
Every skill is either user-invoked or model-invoked:
line 7
Very specific:
Follow writing-for-agents: the top of SKILL.md carries the steps.
line 16
Distinctive:
No em dashes in prose.
line 20

≈ “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.

Keep what fits and drop the rest. Your version, ready for your agents.

Files

AGENTS.md

Conventions for this repo

Skills live flat in skills/<name>/SKILL.md, so every installer (Codex plugins, the skills CLI, copying by hand) finds them.

Invocation

Every skill is either user-invoked or model-invoked:

User-invoked: reachable only when the human types it. Set disable-model-invocation: true in SKILL.md and policy.allow_implicit_invocation: false in agents/openai.yaml. The description is one human-facing line with no trigger phrases.
Model-invoked: the default. No disable-model-invocation, no policy block. The description is model-facing and carries the trigger branches.

Keep the two harness settings in sync; node scripts/check.mjs fails when they drift. A user-invoked skill can load a model-invoked one ("Load the x skill: the Skill tool, or read its SKILL.md") but can never fire another user-invoked skill: when a step needs one, tell the human to type it.

Writing a skill

Follow writing-for-agents: the top of SKILL.md carries the steps; reference that only some branches need goes in a sibling file behind a pointer. Every step ends on a completion criterion that can be checked.
Positive phrasing. Say the behavior you want, not the one you forbid.
Section headings in BRAND.md are an interface: other skills look them up. Rename one only by updating every skill that names it, and the template in skills/brand-context/BRAND-TEMPLATE.md.
Numbers come from scripts or cited sources. A new rule of thumb says where it is from, or is labeled practice.
No em dashes in prose.

Keeping things in step

Adding, renaming, or removing a skill: update .claude-plugin/plugin.json, the README tables (English and Thai), and skills/ask-designer/SKILL.md, so the router never lies.
skills/brand-system/templates/ are copies of the worked example in examples/morning-mist/brand/. Edit the example, then re-copy tokens.json (as tokens.template.json) and brand-board.html.
The Thai font table in design-craft/thai-typography.md is dated. To refresh it, repeat the method in the file.

Before a pull request

node scripts/check.mjs
claude plugin validate .

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.

Check your setup, free