AugmentClaude

LPM Config

Create and edit YAML configuration files for LPM projects, services, and actions.

Installation

  1. Make sure Claude is on your device and in your terminal.

    Skills load from ~/.claude/skills/ when Claude Code starts up — so you need it on your machine first. If you don't have it yet, install it once with the command below, then run claude in any terminal to verify.

    One-time setup
    npm i -g @anthropic-ai/claude-code

    Already have it? Skip ahead.

  2. Paste into Claude Code or into your terminal.

    This copies the whole skill folder into ~/.claude/skills/lpm-config-gug007/ — the SKILL.md plus any scripts, reference docs, or templates the skill ships with. Safe default: works for every skill.

    Faster alternative (instruction-only skills)

    Skips the clone and grabs only the SKILL.md file. Don't use this if the skill ships Python scripts, reference markdowns, or asset templates — they won't be downloaded and the skill will fail when it tries to load them.

    Quick install (SKILL.md only)
    Sign up to copy
  3. Restart Claude Code.

    Quit and reopen Claude Code (or any other agent that loads from ~/.claude/skills/). New skills are picked up on startup.

  4. Just ask Claude.

    Skills auto-activate when your request matches the skill's description — no slash command needed. Trigger phrases live in the skill's own frontmatter; you can read them in the “What this skill does” section above.

Prefer to read the source first? Open on GitHub.

When Claude uses it

Create and edit lpm YAML configs for projects, services, actions, terminals, profiles, SSH projects, and shared config layers. Use when the user explicitly asks to configure lpm, add an lpm service/action/button/terminal, edit `.lpm.yml`, or manage lpm templates or global config. For operating running projects, use `lpm-cli`.

What this skill does

Configure lpm

Create, modify, and delete lpm config files on macOS. Prefer the lpm config commands for target resolution and validation because they use the same configuration model as the app.

Load only the relevant reference

Resolve the target

Honor an explicit project or layer from the user. Otherwise run:

lpm config resolve --cwd . --json

Interpret the result:

  • One match: edit the returned path without asking for confirmation.
  • Multiple matches: ask which candidate to use.
  • No match: offer to create a project for the current directory or edit one of the returned available projects.

Use these explicit layer overrides:

  • “globally” or “all my projects” → ~/.lpm/global.yml
  • “share with the team”, “check it in”, or “for everyone” → <root>/.lpm.yml
  • “template” or “reuse across projects” → ~/.lpm/templates/<name>.yml

When lpm config resolve is unavailable, inspect the YAML project files directly. Match the current directory against expanded root paths by path components; the deepest root wins. Do not parse YAML with line-oriented tools.

Choose the shape

IntentConfig shape
Long-running process started with the projectservices
One-shot command or buttonactions
Persistent interactive shellterminals
Start prerequisites firstservice dependsOn
Ask for parametersaction inputs
Default action plus alternativesparent cmd plus nested actions
Menu with no defaultnested actions without parent cmd
Hidden command with completion notificationtype: background
Send command into the focused terminaltype: command
Reused visible panetype: terminal plus reuse: true
Team-shared config<root>/.lpm.yml
Reusable action settemplate plus extends
Remote executionproject ssh block

Ask only when the requested shape is genuinely ambiguous. For “button with options,” distinguish a split button, dropdown, and input prompt. For a terminal action, distinguish one reused pane from a fresh pane each run.

Apply the change

For a new project, inspect repo signals such as package.json, Makefile, docker-compose.yml, Procfile, and mise.toml. Detect long-running services separately from one-shot test, lint, build, migration, and deploy actions.

For an existing file, preserve unrelated fields and formatting where practical. Use shorthand only when an entry needs no options. Set confirm: true for destructive operations.

Confirm only before deleting a config file or overwriting an existing file during creation. Do not confirm an unambiguous target.

After every write, run:

lpm config validate <path> --json

Fix every reported error before finishing. Report warnings that affect the requested behavior. If the command is unavailable, follow the fallback checklist in references/validation.md.

Related skills