do-roast

Generates character-voice code roasts grounded in evidence dossiers from static, docs, and web analysis.

2|Updated Apr 4, 2026
One-click install
npx skills add https://github.com/DoctorMozg/claude-pipelines --skill do-roast-doctormozg
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: do-roast
Source: https://github.com/DoctorMozg/claude-pipelines/tree/main/plugins/mz-funny/skills/do-roast
Command: npx skills add https://github.com/DoctorMozg/claude-pipelines --skill do-roast-doctormozg

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Teams want humorous, personality-driven code feedback without the roast inventing fake bugs or attacking real people. This Skill produces a character-voice roast where every joke cites a numbered, evidence-backed finding. ## Core Features & Use Cases - Evidence-anchored roasting: Static structure analysis, docs-coherence checks, and web best-practice research produce a numbered findings dossier; the persona may embellish tone but never invent substance. - Seven persona voices: Choose from caveman, wh40k-ork, pirate, viking, dwarf, drill-sergeant, or yoda, each dispatched as a dedicated agent. - Safety guardrails: Hard prohibition on git blame, author names, and real-person references, plus a citation-density check that flags fabrication drift. - Use Case: After a rough sprint, run a pirate-voice roast of the src/auth/ module to give the team a laugh while still surfacing real issues like a 142-line function and a README claim the code contradicts. ## Quick Start Ask the assistant to roast the src/auth directory as a pirate using the do-roast skill.

Frequently Asked Questions about do-roast

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I roast my code with a character voice?

Invoke the skill with a persona and a target, such as roasting a directory, branch diff, or pasted code snippet. Available personas include caveman, pirate, viking, dwarf, drill-sergeant, yoda, and wh40k-ork.

What makes this code roast different from a normal code review?

The roast renders findings in a comedic persona voice, but every line must cite a numbered finding from a real analysis dossier. For serious review without humor, use a dedicated review or audit skill instead.

Can the roast invent bugs or make fun of specific developers?

No. Personas may embellish tone but cannot invent findings, file paths, or bugs not in the dossier. Git blame, author names, and any real-person references are hard-banned from both analysis and output.

What targets can I roast with this skill?

Targets resolve through a ladder: file or directory paths, a branch diff via --diff or branch: syntax, raw quoted text or stdin, or natural-language descriptions resolved by search. Unresolvable targets halt with a blocked status.

Why did my roast come back as a skeleton or with concerns?

If the persona agent returns empty output twice, the skill builds a skeleton roast from the dossier's roastability notes and marks the run DONE_WITH_CONCERNS. Low citation density relative to the finding count also triggers a concerns flag.