What problem does it solve?
Skills installed from the huashu ecosystem (huashu-*, nuwa, darwin, freud, seedance, tramstop, dukou) are continuously iterated, and locally installed copies silently fall behind. This Skill detects outdated installations so users never run stale versions without knowing it.
Core Features & Use Cases
- Full-directory scan: Runs a Python script that scans the skills directory, identifies huashu-family skills via
.git origin or .huashu-skill-meta.json, and compares local commits against remote HEAD on GitHub.
- Three-state report: Outputs each skill as up-to-date, outdated (with the exact update command), or unknown, and exits with code 1 when updates exist so it can plug into cron jobs.
- Guided update workflow: Distinguishes git-clone installs (
git pull --ff-only) from copy installs (re-clone and overwrite the subdirectory), warns before overwriting local modifications, and refreshes last_checked timestamps with --mark-checked.
- Use Case: A user says "check my huashu skills for updates" once a month; the agent runs the script, translates the report into plain language, lists which skills are behind by how many commits, and asks which ones to update.
Quick Start
Ask your AI agent to check whether all installed huashu skills are up to date and list any available updates.