DevNullLtd/Yuzu / .claude/skills/adversarial-review
adversarial-review
VerifiedRun an adversarial two-phase code review of a change with TWO independent reviewers — Claude and Codex — who review alone, then cross-examine each other's findings, then Claude synthesizes a single weighted verdict. Use when the user says "/adversarial-review", "adversarial review", "review this with Codex", "get Codex to review", "two-reviewer review", "cross-examine this PR", or wants a second independent model to grade a change before merge.
adversarial-review
Two independent reviewers grade the same change against a shared set of anchor documents, then
cross-examine each other before anything is reported. Reviewer A is Claude (run as an isolated
subagent so its Phase-1 context can't see the peer). Reviewer B is Codex (the codex CLI, run via
codex exec). You — the Claude running this skill — are the orchestrator, not a reviewer: you set
up the run, enforce the barrier between phases, and synthesize the two Phase-2 reports into one verdict.
This catches what a single reviewer misses: each model has different blind spots, and the Phase-2 cross-examination forces every finding to survive an adversary who is rewarded for refuting it. It is read-and-report only — no GitHub posting, no push, no file edits.
It is heavier than /code-review (two models, four invocations, real compiles). Reach for it on
high-stakes changes — security-surface PRs, frozen-contract work, anything you want a second model to
sign off before merge — not routine diffs.
Bundled files (in this skill dir)
review-prompt.md— the verbatim two-phase reviewer prompt with{{TOKENS}}placeholders. Do not paraphrase it — both reviewers must receive the identical body so "BLOCK" means the same thing.run-codex-reviewer.sh— renders the prompt and runscodex execwith the right sandbox/cwd/writable flags. Used for the Codex side of each phase.--helpdocuments every flag.
Usage
/adversarial-review <target>
<target> is one of:
- a PR number —
/adversarial-review 1220(resolve head SHA + base viagh) - a commit range —
/adversarial-review dev..HEADor87b3795..53c94cf working— review the uncommitted working tree (staged + unstaged + untracked)- empty — default to
dev..HEAD; if that's empty on branchdev, ask whetherorigin/dev..HEADis meant
Prerequisites (check once, up front)
codexon PATH and authenticated —codex login statusshould say "Logged in". If not, tell the operator to runcodex loginthemselves (interactive; can't be done for them) and stop.ghauthenticated if<target>is a PR number.- A configured build dir for the platform under review, if you want Codex's empiricism (compiles/tests)
to succeed under the
workspace-writesandbox — it can't run network-dependentmeson setupfrom a cold tree. See "Sandbox levels" below.
Step 0 — Resolve the run (orchestrator, inline)
- TARGET. Turn
<target>into the full descriptor the reviewers need: repo path, head SHA, and diff range. For a PR:gh pr view <n> --json headRefOid,baseRefName,title→ head SHA; diff range is<merge-base>..<head>. Compose a one-line TARGET string, e.g.PR #1220 "<title>", repo /Users/<user>/Yuzu, head 53c94cf, diff 87b3795..53c94cf. - ANCHORS. The authoritative ground truth severity is graded against. Pick them from the changed
files via the CLAUDE.md routing table — e.g. a Guardian change pulls in
docs/yuzu-guardian-design-v1.1.md §24+ any frozen contract; an auth change pulls indocs/auth-architecture.md; everything getsCLAUDE.md+docs/agentic-first-principle.md(A1–A4). If the anchor set isn't obvious, ask the operator — a wrong anchor list silently mis-grades every blocking call. Format as a markdown bullet list (passed verbatim to both reviewers). - REVIEW_DIR. A shared scratch dir both reviewers read+write, e.g.
/tmp/yuzu-advrev-<short-slug>(slug from PR number or range).mkdir -pit. Keep it OUTSIDE the repo so the git tree stays clean —run-codex-reviewer.shpasses it to Codex via--add-dirso theworkspace-writesandbox can still write there. - Write a one-paragraph scope note to
REVIEW_DIR/TARGET.md(target + anchors + range) so both sides and any later reader share the framing.
State the resolved TARGET, ANCHORS, and REVIEW_DIR back to the operator before launching.
Step 1 — Phase 1: independent review (parallel, then barrier)
Launch both reviewers concurrently, in a single message with two tool calls. The standard pair is Claude + Codex; an orchestrating skill may substitute Kimi through the bundled adapter below as long as both reviewers receive the same prompt body, anchors, target, and phase barrier:
- Claude side —
Agenttool,subagent_type: general-purpose(it needs Bash/Read/Grep for empiricism). Prompt =review-prompt.mdrendered withSELF=claude,PEER=codex,PHASE=1, and the TARGET / REPO / REVIEW_DIR / ANCHORS resolved in Step 0. The subagent's isolated context IS the guarantee it can't peek at Codex's review. Tell it to writeREVIEW_DIR/claude.phase1.mdand return a one-paragraph summary + VERDICT. - Codex side —
Bash,run_in_background: true(Codex compiles/tests; it can run for many minutes — don't block the turn):
(Addbash .claude/skills/adversarial-review/run-codex-reviewer.sh \ --phase 1 --review-dir "$REVIEW_DIR" --repo "$REPO" \ --target "$TARGET" --anchors "$ANCHORS_MARKDOWN"--model <name>to pin a Codex model, or--sandbox danger-full-accessif empiricism needs it.)
BARRIER. Do not start Phase 2 until both REVIEW_DIR/claude.phase1.md and
REVIEW_DIR/codex.phase1.md exist and are non-empty. Poll the background Codex task / re-check the
files. If Codex's phase file is missing after it exits, read REVIEW_DIR/codex.phase1.summary.md to see
what happened (auth failure, sandbox-denied build, etc.) and surface it rather than proceeding with a
one-sided review.
Kimi adapter and static-first mode
run-kimi-reviewer.sh writes kimi.phaseN.md and kimi.phaseN.summary.md using the same prompt and
schema. Its default is static and denies filesystem/search/write tools plus Bash. Trusted dynamic
mode requires
--dynamic --i-trust-this-input;
--sandboxed-dynamic routes Bash through the no-network Docker boundary and fails closed to static
when that boundary is unavailable. When Kimi replaces Claude, use kimi.phase1.md/kimi.phase2.md
at the barriers and use a fresh Kimi invocation for each phase. Kimi 0.17 restricts writes to its
workspace, so restricted modes require an excluded review directory inside the detached worktree;
copy that evidence to external scratch before removing the worktree. Restricted Kimi receives a
bounded materialized bundle (diff, relevant source context, anchors, prior reviews, and orchestrator
evidence) in its prompt; all filesystem/search/write tools and direct Bash are denied. Sandboxed
dynamic mode exposes only one fixed Docker collector through a fail-closed hook and returns its output
in the denied tool result.
An orchestrator may deliberately make Phase 1 a static safety gate before executing PR-controlled
code. Pass Codex --static-only --sandbox read-only, leave Kimi static, record the static synthesis,
and begin empirical work only after the trust gate. This explicit mode overrides the normal Phase-1
empiricism requirement; Phase 2 must ingest the separately recorded dynamic evidence and disclose any
remaining gap.
Step 2 — Phase 2: cross-examination (parallel, then barrier)
Re-invoke both reviewers — fresh each, because all cross-phase state is on disk in REVIEW_DIR. Same two-tool-calls-in-one-message pattern:
- Claude side — a NEW
Agentcall (not a continuation), prompt rendered withSELF=claude,PEER=codex,PHASE=2. It readscodex.phase1.md+ its ownclaude.phase1.md, cross-examines, and writesclaude.phase2.md. - Codex side —
run-codex-reviewer.sh ... --phase 2 ...(againrun_in_background: true).
BARRIER. Wait for both *.phase2.md files.
Step 3 — Synthesis (orchestrator, inline — this is YOUR job, not the reviewers')
Read claude.phase2.md and codex.phase2.md. Neither reviewer wrote a merged verdict; you do. Apply
this weighting — higher signal first:
- Dedupe. Collapse
agrees-with-mine/confirmed-independentlycross-links into single findings. A defect both reviewers reached independently is the strongest class — rank it top. - Weight by provenance.
compiled/test-runevidence outranksstatic-read. A finding one reviewer reproduced empirically and the other only argued from reading is graded on the empirical leg. - Weight by anchor.
contract(cites an ANCHOR section) outranksjudgment. A judgment-only BLOCK is reported but flagged as taste, not contract. - Adjudicate disagreements. Where they split on severity (
disagrees) or one called the otherfalse-positive/unfair, go to the cited code/anchor yourself and rule — show the line that decides it. Don't average; decide. - Surface coverage gaps. Cross the two FILES/COVERAGE lists. Any axis or file neither reviewer went deep on is an explicit gap in the final report — don't let parallel breadth hide a shared blind spot.
- Flag unresolved. Anything left
not-verifiedby both, or where you genuinely can't adjudicate, is reported as an open question, not silently dropped.
Produce the final report (and write a copy to REVIEW_DIR/SYNTHESIS.md):
- Verdict:
BLOCK(any surviving CRITICAL/HIGH) orPASS, one sentence. - Consolidated findings, ranked, each with severity, the winning provenance, anchor-or-judgment, and the minimal fix. Note for each whether it was found by both / one / surfaced only in cross-exam.
- Adjudicated disagreements — what split, how you ruled, the deciding evidence.
- Coverage gaps & open questions.
- What each reviewer ran (compiles/tests/CI) so the empiricism is auditable.
Guardrails
- Read-and-report only. Neither reviewer nor the orchestrator posts to GitHub, comments, pushes, or
edits source. The prompt forbids it; don't add a "post the findings" step unless the operator asks
after seeing the synthesis (then it's
/code-review --commentterritory, a separate action). - Identical prompt body. The only per-reviewer differences are SELF/PEER/PHASE and the shared TARGET/ANCHORS/REVIEW_DIR. Never hand one reviewer a different question than the other.
- Independence is load-bearing. The Claude reviewer is always a subagent, never you-the-orchestrator — you've already seen both sides, so you can't produce an uncontaminated Phase-1.
- Sandbox levels (Codex).
workspace-write(default) lets Codex compile/test in the repo and write to REVIEW_DIR, but denies network — fine when a build dir already exists. If empiricism needs network or system access (coldmeson setup, package installs), pass--sandbox danger-full-access, and only on a host the operator trusts. State which level you used in the synthesis so static-vs-empirical weighting is honest. - One-sided is not a review. If Codex can't run (auth, sandbox, crash) and only
claude.phase*.mdexists, say so plainly and offer a single-reviewer fallback — don't dress a solo Claude review up as adversarial.
Codex side — what run-codex-reviewer.sh does
Renders review-prompt.md → runs codex exec --cd <repo> --sandbox <level> --add-dir <review_dir> --skip-git-repo-check --ephemeral -o <summary> with the prompt on stdin. Ephemeral + on-disk state
means Phase 1 and Phase 2 are independent process invocations — no Codex session to resume, matching the
protocol's disk-barrier model. Codex writes its full review to REVIEW_DIR/codex.phaseN.md itself; the
-o summary file is a fallback for diagnosing a run that didn't produce the phase file.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis. | Stars: 466k | Updated | Claude CodeCodex | Unverified | Install |
| reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis. | Stars: 466k | Updated | Claude CodeCodex | Unverified | Install |
| diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness. | Stars: 319k | Updated | Claude CodeCodex | Verified | Install |
| dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies | Stars: 319k | Updated | Claude CodeCodex | Verified | Install |