Skip to content
Back to skills

Apm Dotfiles

ASecurity

Use when deciding where Stuart's reusable agent customization belongs: public dotfiles vs private overlay, dot-agent vs dot-skill, APM target compatibility, or plugin/config-fragment structure. Not for repo Rust details.

  • 3 stars
  • 0 votes
  • 0 copies
  • 3 views
  • Added August 31, 2026
toolsrusttesting

Works with

  • mcp

Security analysis

A100/100

Scanned October 1, 2026

npx -y skills add sneivandt/dotfiles --skill apm-dotfiles --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Apm Dotfiles?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Apm Dotfiles
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/sneivandt-apm-dotfiles/badge)](https://www.skillsdirectory.com/skills/sneivandt-apm-dotfiles)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: apm-dotfiles
description: >
  Use when deciding where Stuart's reusable agent customization belongs:
  public dotfiles vs private overlay, dot-agent vs dot-skill, APM target
  compatibility, or plugin/config-fragment structure. Not for repo Rust details.
---

# APM Dotfiles

## Place the change

| Content | Owner |
|---|---|
| reusable interaction or coding workflow | `dot-agent` |
| skill/plugin authoring and maintenance | `dot-skill` |
| repository-specific guidance | that repository's `.agents/skills/` |
| work, host, employer, or private guidance | private overlay plugin |
| model/reasoning/UI defaults | public `conf/agent-settings.toml` |
| skills, hooks, prompts, instructions, MCP | APM plugin or fragment |

Prefer one narrow skill per recurring behavior. Update an existing skill instead
of adding a near-duplicate.

## Plugin rules

- Use native layout: `apm.yml`, `includes: auto`, and `.apm/<primitive>/`.
- Keep local package names short and `dot-*`.
- Reference local packages with forward slashes.
- Public files contain no secrets, private URLs, or employer-confidential data.
- Private overlays add fragments/plugins through overlay symlink config and
  reference credentials through environment variables.
- Keep self-defined MCP servers direct unless transitive trust is intentional.

## Target compatibility

- Top-level `targets` selects default harnesses.
- A dependency-level `targets` list narrows the entire package.
- Split packages when primitives need different target sets.
- Cowork remains an experimental, skills-only target. A Cowork-targeted
  package's skills must work without its MCP servers, hooks, prompts, agents,
  commands, or instructions.
- Dotfiles enables Cowork's APM feature and honors APM's configured path, then
  follows lockfile target subsets during ACL-safe file reconciliation.
  Dependency filters are authoritative.

## Verify

Use the dotfiles checkout's `docs/TESTING.md` for scoped checks and previews,
and its `ai-tooling-apm` skill for implementation details. Skill-text-only edits
need frontmatter, link, and routing review, not a live install or deployment.
For behavior changes, prefer current-source checks with isolated homes and
synthetic overlays; a wrapper may bootstrap or run an older installed binary.
An authorized preview with a real overlay can execute its cooperative
`--check`/`--dryrun` scripts. Inspect trusted scripts first and keep private paths
and output out of public reports.

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.