agentleFS
fmindfmind/dot/.agents/skills/dot-skills/SKILL.md

dot-skills

Skill13 starsChanged 5 days agoMITView on GitHub

Maintain fmind/dot skill packages, catalog budgets, package contracts, and installed links.

1 other project uses a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

3 rules. About 0 are common, 1 is distinctive, and 2 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈1

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 2

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Locate references; rename atomically: use rg to find the old name and path across both catalogs, docs, host configuration, and dot/.

    line 27
  2. Distinctive:

    Repoint all links at once: update inbound links and the global link declaration together.

    line 28
  3. Very specific:

    Keep ~/.agents/skills/ a real shared directory: never exact_ or a whole-catalog symlink.

    line 33

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill 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
SKILL.md

Maintain Dot Skills

Maintain first-party skills and their chezmoi links. skillify owns authoring; its package rules define package shape, resources, and standalone portability.

Workflow

  1. Choose the owner: inspect git diff and git diff --cached, read neighboring descriptions, and apply skillify's admission rule. Reusable skills belong in skills/; repository procedures belong in .agents/skills/.
  2. Register additions: each scope (AGENTS.md plus skill names, descriptions, and paths) must stay below 5,000 estimated tokens per context budgets; mise run check:skills enforces both scopes from source. Check headroom first with dot agent context --source . --project . and reduce overhead without losing distinctive triggers when full; an addition with under about 60 tokens of global headroom must free an index entry (merge or remove a skill) in the same change. Each global skill needs dot_agents/skills/symlink_<name>.tmpl containing {{ .chezmoi.sourceDir }}/skills/<name>.
  3. Connect resources: use metadata.kind as the single connector, task, or collection tag. Keep a single procedure in the root; preserve substantial optional modes as guides at references/<name>.md or references/<name>/GUIDE.md, with their owned resources. Run mise run format:skills to generate parent routing indexes. Never nest SKILL.md. Update project AGENTS.md when workflow ownership changes; setup, usage, and the short task reference belong in README.md.
  4. Validate: run mise run check:skills and formatting checks for edited files. For installation changes, also target uv run --frozen --project dot pytest -q dot/tests/test_skill_install_links.py. Follow project AGENTS.md for broader qualification; prose-only changes do not require the full Python suite or build. Use git-worktree when formatters could alter unrelated work. dot agent context --check without --source also measures independently installed packages in the shared roots. Review dot agent context --source . --project . --details when descriptions or guides change; its estimates are diagnostic.
  5. Exercise changed guidance per skillify; when host integration changes, check the affected host separately, since catalog checks do not prove discovery or selection.

Rename or remove

  1. Locate references; rename atomically: use rg to find the old name and path across both catalogs, docs, host configuration, and dot/. For a rename, change the directory, frontmatter name, provenance path, and update date together; preserve needed guidance when consolidating packages.
  2. Repoint all links at once: update inbound links and the global link declaration together. Preserve historical records, then search again and validate as above.
  3. Retire installed links with .chezmoiremove: list the retired .agents/skills/<name> targets there so apply deletes them on every machine, per the chezmoi skill. Use the installed-link recovery guide for manual cleanup or source relocation.

Boundaries

  • Keep ~/.agents/skills/ a real shared directory: never exact_ or a whole-catalog symlink. Independently installed packages remain outside this repository; names must be unique, and apply rejects a name owned elsewhere: choose another name instead of overwriting that package.
  • Know each check's coverage: mise run check:skills validates both first-party roots; gh skill publish --dry-run skills alone covers only the global directory. Sibling references are allowed here; standalone publication needs the package rules' portability check.

Documentation

  • skill_contracts.py enforces package shape, resources, guide indexes, and context budgets; contract tests cover it and chezmoi link registration; lychee checks Markdown links.
  • Companion skills: agent-project (host discovery), repository-docs (documentation ownership).

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.