メニュー
Compatibility alias for stricter Claude review; prefer claude-review with --adversarial for new use.
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
SOC 職業分類に基づく
Diagnose whether Claude Code is installed and authenticated for this local Codex plugin without making a billable Claude call.
Run a structured, read-only Claude Code review over the current git diff or a branch diff.
Delegate a foreground read-only or write-capable task to Claude Code from Codex using the local companion script.
Configure or explain the optional Codex Stop hook that can run a read-only Claude review when a turn ends.
Cancel a running Claude Code companion job by sending process signals through the plugin job index.
Analyze plugin-owned Claude Code job logs for permission prompts and export reviewed allowedTools arguments.
| name | claude-adversarial-review |
| description | Compatibility alias for stricter Claude review; prefer claude-review with --adversarial for new use. |
Use claude-review with --adversarial for new stricter reviews. This compatibility skill remains available for parity with earlier plugin versions.
Use the companion script from the installed plugin root. If the current working
directory is not this plugin checkout, resolve the script relative to this skill
file, for example ../../scripts/claude-companion.mjs, and pass --cwd <target-repo> for the repository being reviewed.
Real adversarial review calls send diffs and review prompts to Claude Code and may spend quota. If the Codex host offers persistent approvals, prefer approving the narrow prefix node scripts/claude-companion.mjs review and running review --adversarial instead of approving a separate adversarial command. If host policy blocks external disclosure, do not bypass it.
Review uncommitted changes:
node scripts/claude-companion.mjs review --adversarial
Review a branch diff:
node scripts/claude-companion.mjs review --adversarial --base main
Legacy compatibility alias:
node scripts/claude-companion.mjs adversarial-review --base main
For machine-readable output:
node scripts/claude-companion.mjs review --adversarial --base main --json
This uses the same schema, diff collection, and JSON parsing as claude-review, but with a stricter prompt that challenges assumptions and looks for concrete failure modes. It always uses Claude Code --permission-mode plan and should not edit files.
It also uses the same single-review diff limit as claude-review: 200000 bytes by default, configurable with --max-diff-bytes <n>.
Use --effort <level> to pass Claude Code effort: low, medium, high, xhigh, or max.
Model guidance:
--model opus for serious adversarial review, merge readiness, or high-risk design validation.--effort low only for smoke tests, cheap sanity checks, or explicitly low-effort requests.