code-roast
Code Roast
Give any codebase a Gordon-Ramsay-style roast: a letter grade, a Hall of Shame, genuine Bright Spots, and a Prescription — all in a shareable Markdown report.
Workflow
Step 1 — Identify the repo root
If the user didn't specify a path, use the current working directory as the repo root. Confirm it is a git repo or a directory with code files. Store it as <repo_root>.
Step 2 — Run the analyzer
cd <skill_dir>/scripts
uv run analyze.py <repo_root> --debug
This emits a JSON object to stdout with 9 shame-category metrics. Copy the full JSON — you'll need it in Step 3.
More from likw99/agent-skills
sync-trending
Monitior technology trends (GitHub, etc.), contextualize them against the user's project, and autonomously verify them through installation and testing. Use when the user asks about trending repositories, new tools, or wants to stay updated on tech relevant to their current work.
9llm-daily
>
4dev-card
>
2trustskills
Use this skill when a user wants a trust decision before installing from a skill URL, marketplace, or GitHub repo. It checks a compact allowlist of trusted distribution channels and returns whether the source should be trusted under the current TrustSkills policy, without drifting into explaining what the skill itself does unless the user asks.
1