agents
CLAUDE.md0 starsChanged 4 months agoMITView on GitHub
This repository is the source of truth for Uke's personal agent assets: skills, plugins, and the metadata needed to distribute or synchronize them. It is not an application repository and currently has no runtime service, database, CI/CD pipeline, or test suite. - skills/ contains personal, distributable skills. Each skill lives in its own directory and has a SKILL.md entry point; keep its scripts, references, and other bundled assets beside it. - plugins/ is reserved for personal plugins created in this…
- Reads credentials
In this repository11 other files it gives its agents.
- AGENTS.mdAGENTS.md
- create-agentsmdSkill
- create-readmeSkill
- create-subagentsSkill
- find-skillsSkill
- github-actions-docsSkill
- github-actions-templatesSkill
- plugin-creatorSkill
- Plugin StructureSkill
- role-creatorSkill
- skill-creatorSkill
The file
We split this CLAUDE.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
- Distinctive: line 32
Do not introduce conventions for future top-level asset types without an explicit repository decision.
- Very specific: line 38
Use Node.js >=22 and pnpm >=10.30.3 <12 as specified in package.json.
- Very specific: line 41
Check only formatting with pnpm run format:check.
- Very specific: line 56
Read the target asset's SKILL.md, plugin manifest, and nearby documentation before modifying it.
- Very specific: line 59
Keep skills focused and self-contained.
- Distinctive: line 62
When changing a user-facing asset or its distribution—for example, adding, removing, renaming, or materially changing a skill or …
- Very specific: line 69
Do not place secrets or user-specific credentials in the repository.
- Very specific: line 74
Use Conventional Commits. The commit-msg hook enforces this through Commitlint.
≈ “Common” is an estimate: at least 1 in 5 of the 26,980 CLAUDE.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
Purpose
This repository is the source of truth for Uke's personal agent assets: skills, plugins, and the metadata needed to distribute or synchronize them. It is not an application repository and currently has no runtime service, database, CI/CD pipeline, or test suite.
Repository boundaries
skills/contains personal, distributable skills. Each skill lives in its own directory and has aSKILL.mdentry point; keep its scripts, references, and other bundled assets beside it.plugins/is reserved for personal plugins created in this repository. Keep each plugin self-contained in its own directory.skills.sh.jsonis the checked-in catalogue configuration for skills exposed through skills.sh. When adding, removing, or renaming a listed personal skill, update the applicable grouping in this file.skills-lock.jsonrecords externally installed development skills and their sources. Treat it as generated metadata: update it only through the tool that manages those installed skills; do not hand-edit hashes.README.mdis the English, reader-facing overview.README.zh_CN.mdis its Simplified Chinese counterpart; keep their user-facing information semantically equivalent..agents/and.claude/are development-support configuration for Codex and Claude Code while they work on this repository. They are not locations for personal distributable assets. Never add a personal skill under.agents/skills/or.claude/; add it underskills/instead. Likewise, personal plugins belong underplugins/.
Do not introduce conventions for future top-level asset types without an explicit repository decision. Ask before creating a new top-level source directory.
Environment and commands
- Use Node.js
>=22and pnpm>=10.30.3 <12as specified inpackage.json. - Install development dependencies with
pnpm install. - Check only formatting with
pnpm run format:check. - Reformat the repository with
pnpm run format. This command writes broadly; do not run it when unrelated working-tree changes must be preserved.
There is no automated test command at present. Validate a changed skill or
plugin using the smallest relevant workflow described by that asset, then run
pnpm run format:check before handing off a change.
pnpm check is intended to aggregate formatting and dependency checks, but it
currently fails because pnpm run deps:check invokes the unavailable command
pnpm peers check. Do not treat this as a validation gate until the script is
repaired; report the failure if it is relevant to the task.
Working conventions
- Read the target asset's
SKILL.md, plugin manifest, and nearby documentation before modifying it. Follow its local instructions when they are more specific than this file. - Keep skills focused and self-contained. Use relative links from
SKILL.mdto bundled references or scripts, and do not depend on files in.agents/or.claude/at runtime. - When changing a user-facing asset or its distribution—for example, adding,
removing, renaming, or materially changing a skill or plugin—check whether
the READMEs need updating. If they do, update both
README.mdandREADME.zh_CN.mdin the same change, including asset lists, installation commands, requirements, and status statements. - Preserve the existing directory structure and avoid unrelated formatting, dependency, or lockfile changes.
- Do not place secrets or user-specific credentials in the repository. Use
environment variables or a local ignored
.envfile when a tool needs them.
Versioning and commits
Use Conventional Commits. The commit-msg hook enforces this through
Commitlint; examples include feat(skills): add pdf reading workflow,
fix(plugins): correct manifest path, and docs: clarify installation.
Changesets are installed but are not part of the current workflow: the
repository currently contains configuration assets rather than versioned,
published workspace packages. Do not create a changeset for skill, plugin, or
repository-configuration changes. Revisit this policy before adding a package
under plugins/* or tools/* that is independently versioned and published.
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.

