メニュー
Run structured code/security review for local, branch, or commit diffs with explicit RED-safe routing.
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
Improve, audit, simplify, rewrite, or migrate prompts, tool descriptions, agent instructions, and prompt stacks for GPT-5.6 Sol or the GPT-5.6 family. Use for outcome-first prompt design, autonomy boundaries, tool routing, PTC, grounding, verbosity, reasoning effort, and prompt evals.
Apply a deep, design-minded engineering workflow for complex work that needs careful planning, iteration, and simplification.
Use when preparing or running a Claude CLI agentic engineering review through claude -p for repository audits, architecture analysis, system design, security review, large refactors, specs, RFCs, or evidence-grounded long-form engineering analysis.
Use when preparing or running a ZCode GLM-5.2 agentic builder workflow through zcode --prompt for fast implementation, iterative code generation, focused fixes, and validation on an existing repository.
Apply adversarial opposite-analysis to plans, specs, architecture, code changes, and claims. Use when the user asks for adversarial review, opposing analysis, contrarian review, red-team reasoning, or Z.ai and MiniMax cross-checks through the Ralph MCP router.
Review and adjudicate Bugbot, Cursor, Seer, and similar automated PR feedback with local evidence before accepting, fixing, or dismissing findings.
| name | autoreview |
| description | Run structured code/security review for local, branch, or commit diffs with explicit RED-safe routing. |
Use this repo-local skill for AI-assisted code review when the reviewed content is safe for the selected reviewer engine.
Canonical implementation for review lives at skills/autoreview/ in this repo.
Run it from the target repository:
python3 skills/autoreview/scripts/autoreview.py --mode auto --dry-run
python3 skills/autoreview/scripts/autoreview.py --mode local --review-pass 1 --review-total 10
python3 skills/autoreview/scripts/autoreview.py --mode branch --base origin/main --review-pass 1 --review-total 10
python3 skills/autoreview/scripts/autoreview.py --mode branch --base origin/main --review-pass 10 --review-total 10
python3 skills/autoreview/scripts/autoreview.py --mode commit --commit HEAD --review-pass 1 --review-total 10
Defaults are Ralph-safe: web search, git fetch, and untracked files are all
opt-in, and sensitive content blocks reviewer execution before any engine call.
Execution is also bounded: choose a pass budget from 1 to 10, run each pass at
most once with --review-pass N --review-total M, stop early on a clean pass,
and stop at pass M even if residual findings remain. Do not commit one finding
at a time or run autoreview until clean.
Keep this local until review passes; promote globally only after explicit user
approval.