Name: roast
Rating: 71.2 (1 reviews)
Author: alirezarezvani

roast

Use when someone asks to roast an idea, pressure-test or stress-test an idea, validate a business idea, "convene the panel", get a brutal second opinion before building something, or says "/roast". Spins up a 5-angle panel (Critic, Champion, Analyst, Investigator, Customer) that attacks the idea from every angle, then a Judge returns one GO / RESHAPE / KILL verdict with the cheapest test to de-risk it.

Quality

88%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Low

Low-risk findings worth noting

The canonical home for this skill is roast in alirezarezvani/claude-skills

Quality

Content

82%Weight 40%Scale 1-5

Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.

A well-structured, highly actionable skill body with concrete prompts, commands, and a verdict template. Its main weaknesses are minor rhetorical padding in the intro and bundle files that are referenced but not actually present, which caps progressive disclosure at 4.

Suggestions

Ship the referenced bundle files (references/*.md, scripts/*.py, assets/*.md) so the in-body links resolve; their absence breaks navigation and limits progressive_disclosure.

Trim the motivational framing in 'What this does' (e.g. 'Claude's default is to agree with you. /roast is the opposite.') to tighten conciseness without losing the adversarial framing.

Add an explicit validate/retry note for the parallel panel step (e.g. re-run a panelist whose output is off-mandate) to close the remaining workflow_clarity validation gap.

Dimension	Reasoning	Score
Conciseness	The body is mostly lean and instruction-led (concrete panelist prompts, bash commands, an output template), with only minor rhetorical padding such as 'Claude's default is to agree with you. /roast is the opposite.' that could be trimmed — efficient but not maximal, so score 4 rather than 5.	4 / 5
Actionability	Fully executable, copy-paste-ready guidance: complete bash invocations with example arguments (brief_builder.py, verdict_synthesizer.py, cheapest_test_designer.py), ready-to-paste panelist prompts, and an exact verdict output template covering the common cases.	5 / 5
Workflow Clarity	A clear three-step sequence (frame → run 5 panelists in parallel → call verdict) with a checkpoint via brief_builder ('tells you if anything critical is still missing before you spend five subagents') and synthesizer veto gates; this is not a destructive operation, so the score-3 cap does not apply, leaving minor validation gaps at 4.	4 / 5
Progressive Disclosure	Well-signaled one-level-deep references with per-file descriptions across References, Scripts, and Assets sections and content appropriately split — good structure — but the referenced directories (references/, scripts/, assets/) and files do not exist on disk, so navigation would break, pulling this below the score-5 anchor.	4 / 5
	Total	17 / 20 Passed

Description

95%Weight 40%Scale 1-5

Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.

A strong description that pairs an explicit, multi-trigger 'Use when' clause with a concrete what-it-does summary of the panel and verdict. Third-person voice and specific actions hold throughout; only specificity leaves minor room for fuller coverage.

Dimension	Reasoning	Score
Specificity	Lists several specific concrete actions — 'Spins up a 5-angle panel', 'attacks the idea from every angle', 'a Judge returns one GO / RESHAPE / KILL verdict', 'the cheapest test to de-risk it' — with only minor coverage gaps, fitting the score-4 anchor rather than the fully comprehensive score-5.	4 / 5
Completeness	Explicitly answers both: 'what' ('Spins up a 5-angle panel... returns one GO / RESHAPE / KILL verdict with the cheapest test') and 'when' (an explicit 'Use when someone asks to roast...' clause with concrete trigger phrases).	5 / 5
Trigger Term Quality	Comprehensive natural trigger coverage including synonyms and the slash command: 'roast an idea, pressure-test or stress-test an idea, validate a business idea, "convene the panel", get a brutal second opinion before building something, or says "/roast"' — phrases a user would naturally say.	5 / 5
Distinctiveness Conflict Risk	Clear niche — adversarial 5-angle panel yielding a single verdict — with distinctive triggers ('convene the panel', '/roast', 'brutal second opinion') that minimize overlap with adjacent skills; voice is correctly third person ('Spins up', 'returns').	5 / 5
	Total	19 / 20 Passed

Validation

81%

Warnings & errors only

Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.

Validation — 13 / 16 Passed

Validation for skill structure

Criteria	Description	Result
frontmatter_unknown_keys	Unknown frontmatter key(s) found; consider removing or moving to metadata	Warning
relative_links	Relative link issues: 5 missing	Warning
referenced_paths_exist	Referenced path issues: 17 missing	Warning

	Total	13 / 16 Passed

Repository: alirezarezvani/claude-skills
Path: .gemini/skills/roast/SKILL.md
Commit: 19392f7

Reviewed: ~3 hours ago

Table of Contents

Discovery Implementation Validation

Is this your skill?

If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.