aurokinaurokin/skills-manager/AGENTS.md

skills-manager

AGENTS.md3 starsChanged 6 months agoMITView on GitHub

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 repository8 other files it gives its agents.
- skills/humanizer/AGENTS.mdAGENTS.md
- agents-mdSkill
- broSkill
- handoffSkill
- html-artifactSkill
- humanizerSkill
- ponytailSkill
- unslopSkill

The file

We split this AGENTS.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

Very specific:
Use bun from cli/; skm shells to an external skills executable when fetching or placing upstream skills.
line 41
Very specific:
Add local skills under skills/.
line 42
Very specific:
Add agent definitions as a complete agent.yaml plus non-empty instructions.md pair.
line 43
Distinctive:
Put private skills and host-specific overrides in a registered overlay root, not this public repo.
line 44
Very specific:
Treat skills/agents-md/SKILL.md as generated.
line 45
Very specific:
Never run the archived custom_agents repo's shared-agents tool.
line 47
Distinctive:
Keep CLAUDE.md as a symlink to this file.
line 50
Very specific:
Run the narrowest relevant test while iterating, then cd cli && bun test before handoff.
line 54

≈ “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

Agent Instructions

Purpose

This repo is the public engine and catalog for Skills Manager (skm).
skm manages local, gated, composed, and upstream skills plus agent definitions.
The external skills CLI remains the fetch/place backend for skm upstream sync and skm deploy.

Commands

Task	Command
Install CLI dependencies	`cd cli && bun install`
Run all CLI tests	`cd cli && bun test`
Run one CLI test	`cd cli && bun test test/<name>.test.ts`
Preview managed placements	`cd cli && bun src/cli.ts plan`
Apply managed placements	`cd cli && bun src/cli.ts apply`
Build the review console	`cd cli && bun src/cli.ts review`
Sync upstream skills	`cd cli && bun src/cli.ts upstream sync`
Deploy a project family	`cd cli && bun src/cli.ts deploy <dir> --family <name> --yes`
Validate the `agents-md` fork	`bash maintenance/test-agents-md.sh`
Refresh the `agents-md` fork	`bash maintenance/sync-agents-md.sh`

Repository Map

Need	Path
Operator guide	`README.md`
CLI reference	`cli/README.md`
Integrated architecture	`docs/skills-manager-design.md`
Decisions	`docs/adr/README.md`
Upstream catalog	`catalog/global-specs.txt`
Project families	`catalog/families.tsv`, `catalog/families/*.txt`
Catalog conventions	`catalog/README.md`
Agent capabilities	`registry/README.md`, `registry/agents.json`
Local skills	`skills/<name>/SKILL.md`
Agent definitions	`agents/<name>/agent.yaml`, `agents/<name>/instructions.md`

Conventions

Use bun from cli/; skm shells to an external skills executable when fetching or placing upstream skills.
Add local skills under skills/; add upstream skills to catalog/global-specs.txt.
Add agent definitions as a complete agent.yaml plus non-empty instructions.md pair.
Put private skills and host-specific overrides in a registered overlay root, not this public repo.
Treat skills/agents-md/SKILL.md as generated; change it only through the maintenance sync.
Declare user-invoked-only skills with disable-model-invocation: true; see ADR 0011.
Never run the archived custom_agents repo's shared-agents tool; skm owns those placements now.
Preserve the ownership boundary: plan/apply state never adopts installs made by upstream sync or deploy.
Preserve user changes in dirty worktrees and never delete foreign placements.
Keep CLAUDE.md as a symlink to this file; do not maintain a divergent copy.

Verification

Run the narrowest relevant test while iterating, then cd cli && bun test before handoff.
For catalog or sync changes, run the targeted upstream/deploy unit and parity tests; never exercise real global sync as a test.
For placement or rendered-artifact changes, inspect skm plan; do not apply to the real home during tests.

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