adversarial-review

Cross-vendor adversarial code review of the current branch. Two different model families (Claude + Codex/GPT) review the diff independently, then try to refute each other's findings; survivors are reported by confidence. Runs from either Claude Code or Codex. Use when the user asks for an adversarial review, a cross-model / second-opinion review, or wants high-confidence findings before merging. Report-only — never auto-applies fixes.

Quality

100%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Passed

No findings from the security scan

Quality

Content

100%Weight 40%Scale 1-3

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

A tight, highly actionable orchestration skill: executable commands, a clearly phased workflow with validation gates, and clean separation of prompts/schemas into referenced sibling files. The only minor caveat is that the referenced prompts/ and schemas/ files are not present in this review bundle to verify on disk.

Dimension	Reasoning	Score
Conciseness	Lean prose that assumes competence; extra text is skill-specific rationale ('self-ratification', 'confident false positives', empirically-verified CLI gotchas like the codex stdin block and claude array-output shape) rather than generic concepts Claude already knows.	3 / 3
Actionability	Provides copy-paste-ready executable commands: the argv-array diff construction, exact `codex exec` and `claude -p` invocations with flags, and `< /dev/null` redirection — fully concrete rather than pseudocode.	3 / 3
Workflow Clarity	Clearly sequenced Preflight → Phase 0-3 with explicit validation checkpoints (missing-CLI check, empty-diff stop, deterministic gates before models, synthesize-and-confirm-no-autofix), and feedback loops via the cross-refute step.	3 / 3
Progressive Disclosure	SKILL.md is an orchestration overview that points one level deep to clearly signaled siblings (`prompts/review.md`, `prompts/refute.md`, `schemas/findings.schema.json`, `schemas/verdicts.schema.json`); content is appropriately split out rather than inlined, and sections are well organized. No bundle files are present to verify the referenced paths on disk.	3 / 3
	Total	12 / 12 Passed

Description

100%Weight 40%Scale 1-3

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, third-person description that states concrete capabilities, includes natural trigger terms, and explicitly pairs the 'what' with a 'Use when' clause. Voice is consistently third person, so no specificity penalty applies.

Dimension	Reasoning	Score
Specificity	Lists multiple concrete actions — 'Two different model families... review the diff independently', 'try to refute each other's findings', 'survivors are reported by confidence' — matching the multiple-specific-actions anchor rather than the partial score-2 anchor.	3 / 3
Completeness	Explicitly answers both what (cross-vendor adversarial review with refute step) and when via the 'Use when the user asks for...' clause, satisfying the explicit-trigger anchor.	3 / 3
Trigger Term Quality	Natural user phrasing is well covered: 'adversarial review', 'cross-model / second-opinion review', 'high-confidence findings before merging' — terms a user would actually say, not jargon.	3 / 3
Distinctiveness Conflict Risk	The cross-vendor refute-then-report-by-confidence niche is distinct from ordinary review skills, with 'Report-only — never auto-applies fixes' further scoping it; unlikely to trigger for the wrong skill.	3 / 3
	Total	12 / 12 Passed

Validation

100%

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 — 16 / 16 Passed

Validation for skill structure

No warnings or errors.

Repository: basicmachines-co/basic-memory
Path: .agents/skills/adversarial-review/SKILL.md
Commit: 60408ad

Reviewed: 11 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.