From orphan-reaper
Scan and kill orphan dev processes (storybook, esbuild, vite, next, etc.) left behind by AI agent sessions. Use when the user mentions slow machine, hot CPU, orphan processes, or asks to clean up dev processes.
How this skill is triggered — by the user, by Claude, or both
Slash command
/orphan-reaper:cleanupThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
You operate the `orphan-reaper` CLI on behalf of the user.
You operate the orphan-reaper CLI on behalf of the user.
Always pass --json to get structured output. Never parse human-readable output.
The CLI ships with this plugin. Resolve it in order:
bash ~/.claude/plugins/marketplaces/tomstack/orphan-reaper/cli.sh (plugin install)orphan-reaper (global npm install)On first invocation, try option 1. If the file doesn't exist, fall back to option 2. Store the working path and reuse it for the rest of the session.
All examples below use $CLI as placeholder. Replace with the resolved path.
| Intent | Command |
|---|---|
| List orphan processes | $CLI scan --json |
| Kill all orphans | $CLI kill --json |
| Kill silently | $CLI kill --quiet --json |
| Show matched patterns | $CLI patterns --json |
| Add custom pattern | $CLI add 'my-custom-dev-server' --json |
| Show version | $CLI version --json |
{"ok": true, ...} with exit code 0{"error": "message"} on stderr with exit code 1scan returns count, total_rss_mb, and orphans arraykill returns killed, freed_rss_mb, and processes arrayscan first to show the user what will be killedkill (unless they said "clean up" or "kill")A process is an orphan if it:
Processes with a TTY are intentional (user running them in a terminal) and are never killed.
Users can add project-specific patterns to ~/.orphan-reaper/patterns.conf.
One regex per line. Lines starting with # are comments.
npx claudepluginhub tombelieber/tomstack --plugin orphan-reaperProvides behavioral guidelines to reduce common LLM coding mistakes, focusing on simplicity, surgical changes, assumption surfacing, and verifiable success criteria.
Searches, retrieves, and installs Agent Skills from prompts.chat registry using MCP tools like search_skills and get_skill. Activates for finding skills, browsing catalogs, or extending Claude.
Creates, edits, and optimizes skills for Claude Code, including drafting, evaluating with test prompts, iterating on performance, and improving skill descriptions for better triggering accuracy.