adversarial-review

Three parallel reviewer agents find contradictions, config gaps, and coverage holes across a project's docs, config, and tests — then fix what you approve. Heavyweight (dispatches parallel agents, can modify files); explicit invocation only. Triggers: /adversarial-review, "adversarial review", "find gaps in my docs/config".

Quality

88%

Does it follow best practices?

Impact

—

No eval scenarios have been run

Securityby

Risky

Do not use without reviewing

Quality

Content

77%

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

This is a well-structured, highly actionable skill that clearly orchestrates a complex multi-agent workflow with proper validation checkpoints and error handling. Its main weaknesses are moderate verbosity from repeated content (the AI instruction file list appears 4 times) and being a monolithic document that could benefit from splitting agent prompt templates into separate files. The workflow design with user checkpoint gating and parallel agent dispatch is excellent.

Suggestions

Extract the repeated list of AI assistant instruction files into a single variable or reference at the top, then refer to it throughout instead of repeating the full list 4 times.

Consider splitting the three agent prompt templates into a separate reference file (e.g., AGENT-PROMPTS.md) to reduce the main skill's length and improve progressive disclosure.

Dimension	Reasoning	Score
Conciseness	The skill is quite long (~300 lines) and contains some repetition (e.g., the full list of AI assistant instruction files is repeated 4 times verbatim). However, most content is genuinely instructional and not explaining concepts Claude already knows. The agent prompt templates are necessarily verbose but could be more concise.	2 / 3
Actionability	The skill provides highly concrete, executable guidance: specific git commands, exact file paths to check, precise agent prompt templates, structured output formats, and clear decision trees. Every step tells Claude exactly what to do, with specific commands and file globs.	3 / 3
Workflow Clarity	The 5-phase workflow is clearly sequenced with explicit checkpoints: Phase 1 discovery → Phase 2 parallel agents → Phase 3 user checkpoint (explicit gate before proceeding) → Phase 4 fix agent → Phase 5 summary. The user checkpoint in Phase 3 is a strong validation gate. Error handling table covers edge cases. The retry logic for failed agents provides a feedback loop.	3 / 3
Progressive Disclosure	The skill is a monolithic document with no references to supporting files. While the content is well-structured with clear headers and phases, the agent prompt templates and detailed discovery instructions could be split into separate reference files. The review-config.md and ARCHITECTURE.md references are good external integration points, but the skill itself is a long single file with no bundle support.	2 / 3
	Total	10 / 12 Passed

Description

100%

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

This is an excellent skill description that concisely communicates what the skill does (parallel adversarial review finding contradictions, gaps, and holes), when to use it (explicit triggers listed), and important operational context (heavyweight, modifies files, requires approval). It uses third-person voice throughout and provides clear, distinctive trigger terms that minimize conflict risk.

Dimension	Reasoning	Score
Specificity	Lists multiple specific concrete actions: 'find contradictions, config gaps, and coverage holes across docs, config, and tests' and 'fix what you approve'. Also describes the mechanism ('three parallel reviewer agents') and operational characteristics ('heavyweight, dispatches parallel agents, can modify files').	3 / 3
Completeness	Clearly answers both 'what' (three parallel agents find contradictions, config gaps, coverage holes and fix approved issues) and 'when' (explicit triggers listed, plus 'explicit invocation only' constraint). The triggers section serves as a clear 'Use when' equivalent.	3 / 3
Trigger Term Quality	Includes explicit trigger terms users would naturally say: '/adversarial-review', 'adversarial review', 'find gaps in my docs/config'. Also includes natural domain terms like 'contradictions', 'config gaps', 'coverage holes', 'docs', 'config', 'tests'.	3 / 3
Distinctiveness Conflict Risk	Highly distinctive with a clear niche: adversarial multi-agent review across docs/config/tests. The explicit invocation requirement and specific trigger commands ('/adversarial-review') make accidental triggering very unlikely. This is unlikely to conflict with simpler doc or config skills.	3 / 3
	Total	12 / 12 Passed

Validation

90%

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 — 10 / 11 Passed

Validation for skill structure

Criteria	Description	Result
frontmatter_unknown_keys	Unknown frontmatter key(s) found; consider removing or moving to metadata	Warning

	Total	10 / 11 Passed

Repository: belchman/claude-skills
Commit: be88d6c

Reviewed: 23 days 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.