siteview-cleanup

/siteview-cleanup

Kill stale site-view servers. Often you end up with 3–6 servers running after restarting Claude or re-launching the UI during a session; this skill clears them.

Usage

/pipecrew:siteview-cleanup [--keep-port=<N>] [--keep-run=<run-id>] [--keep-latest] [--all] [--dry-run]

Default: --dry-run — prints what WOULD be killed and exits without touching any process. Opt in to actually kill with one of the explicit flags.

Flags:

• --keep-port=<N> — preserve the server on port N; kill the rest.
• --keep-run=<run-id> — preserve all servers serving this run-id; kill the rest.
• --keep-latest — keep the server on the highest-numbered port; kill the rest.
• --all — kill every PipeCrew site-view server found.
• --dry-run (default if no kill flag set) — preview only.

Instructions

When the user invokes /siteview-cleanup:

1. Run the cleanup script via Bash, passing through the user's flags verbatim:

   node {plugin_dir}/skills/siteview-cleanup/cleanup.js [user's flags]
   

2. Relay the output to the user verbatim — it's a pre-formatted report.
3. If the user didn't specify a kill flag, remind them the default is dry-run and suggest the best flag for their intent:
   • Cleaning up after a session → --keep-latest
   • Keeping one specific run alive → --keep-run=<run-id>
   • Wiping the slate → --all

Safety

• The script ONLY kills processes whose port responds with a PipeCrew site-view shape. Non-site-view processes on the scan range are ignored.
• --dry-run is the default — there's no way to accidentally kill by invoking without flags.
• After kill, the script waits 200 ms and re-probes each killed port to confirm it's gone.

Full reference

Flag semantics + exit codes documented in {plugin_dir}/skills/siteview-cleanup/cleanup.js header.