spectre-code-review

name	spectre-code_review
description	Run final adversarial review through a pinned high-effort opposing runtime with one native fallback. Use at the final boundary to falsify correctness, safety, production readiness, and requirement delivery. Do not use for partial checkpoints or implementation; review only.
user-invocable	true

code_review

Purpose

Try to prove the work wrong, unsafe, unreachable, or unable to meet requirements. Accept correctness only if evidence survives adversarial scrutiny; never defend the implementation.

Inputs

• $ARGUMENTS: feature/root or descendant artifact, focus, diff/base range, source-plan path, immutable BASE_SHA/HEAD_SHA/DIFF_SHA256 tuple, and --orchestrated.
• Scope: completed changes, changed files, direct dependencies/importers, tests, and requirements/ACs. Source requirements from matching tasks.json parents, else an explicit source-plan path before plan.md, then request+diff. execute.md only locates tasks.

Working Set

• Resolve one managed FEATURE_ROOT for this work from explicit/current-thread evidence only (physical directory wins; never branch/recency/lifecycle/scans). If none is confirmed, including when the candidate path is occupied, standalone MUST first load and follow Skill(spectre-feature-root) through DONE; orchestrated calls escalate. Keep writes beneath it and pass it unchanged.
• Repair stale feature/root metadata in artifacts this review touches.
• REVIEW_REPORT={FEATURE_ROOT}/reviews/comprehensive_code_review.md; replace it for the current candidate. Git history preserves prior decisions without accumulating sibling reports.
• Late-bind diff/base and tuple, changed files, requirement/AC paths and task ids, dependencies/importers, tests, and exclusions. Reviewers read diffs/task graphs.

Outputs + DONE

REVIEW_REPORT contains:

0. title, Feature:, Feature Root:;
1. Scope Boundary — completed work, diff/base, tuple, requirements, files/dependencies/tests, exclusions;
2. Verdict — BLOCKED for CRITICAL/HIGH, PASS WITH FINDINGS for MEDIUM/LOW only, else CLEAN;
3. Findings — # | Severity | Lens | Location | Evidence / Reproduction | Impact | Finding Fingerprint | Invariant Family | Smallest Fix, severity-ordered or No findings;
4. Coverage Record — inspected/unverified areas and reasons;
5. Requirement Delivery Coverage — Requirement/AC | Status | Consumer/outcome evidence | Gap/Finding for every applicable item;
6. Scope and Dead-Path Audit — separate tables for scope creep, dead computations/orphaned outputs, old active paths, and duplicate data sources; empty categories say None found;
7. Prioritized Actions — ordered remediation or None;
8. Review Metadata — ISO8601 timestamp, Review Mode: final, runtime/model/effort/route, and fallback reason.

DONE when schema validation passes; scope/tuple are unchanged; every requirement/AC has one evidence-backed status; audit and route metadata exist; findings satisfy identity/evidence rules; and only the report changed.

Method / guardrails

1. Use one external-first review route from repo root; do not probe alternatives. Codex primary runs: claude -p --model opus --effort high --permission-mode dontAsk --allowedTools "Read,Grep,Glob,LS,Bash(mkdir -p *),Bash(git diff *),Bash(git show *),Bash(git status *),Bash(git rev-parse *),Write" --output-format text "$REVIEW_PROMPT" Claude primary runs: codex exec -C "$PWD" -m gpt-5.6-sol -c 'model_reasoning_effort="high"' -s workspace-write "$REVIEW_PROMPT" Record route metadata as Claude Code|opus|high|Codex -> Claude Code or Codex|gpt-5.6-sol|high|Claude Code -> Codex. Keep a 20-minute launcher-side poll limit; do not pass duration guidance to the reviewer. Quiet output is not failure. The reviewer writes only REVIEW_REPORT. Primary semantic self-review is prohibited.
2. Fallback once. Missing/non-zero opposing CLI, report-write failure, or unusable review permits one clean-context @spectre_reviewer under the same manifest, contract, and schema. It returns the report in-thread for unchanged persistence. Record native-subagent|runtime-native|inherited|native-fallback plus reason. A usable review ends semantic review.
3. Falsify; do not confirm. Actively seek counterexamples, broken invariants, failure paths, false-positive tests, unreachable outcomes, and evidence contradicting claimed delivery. Cover correctness; regression/integration; security; performance/reliability; materially avoidable overengineering; test adequacy; requirement reachability; scope/dead paths. Stay within completed work and direct blast radius; omit style, praise, speculation, subjective preferences, and requirements from another scope.
4. Require evidence. Severity is CRITICAL|HIGH|MEDIUM|LOW: CRITICAL means exploit/data loss/corruption/privilege bypass/core failure; HIGH a concrete user-facing defect, serious security weakness, or demonstrated hot-path/reliability failure; MEDIUM a localized defect or material maintainability/performance/test risk; LOW a small actionable issue. Each finding cites file:line, violated behavior, evidence, impact, smallest scope-safe fix, finding_fingerprint=sha256(requirement anchor + primary symbol/boundary + normalized observable failure), and invariant_family=sha256(requirement anchor + normalized violated invariant + lifecycle/data-flow boundary). CRITICAL/HIGH also requires the normalized invariant and reproducible failure/exploit/performance path; otherwise downgrade or omit.
5. Prove delivery. Assign each requirement/AC one of Delivered|Partial|Dead|Missing. Delivered requires compact usage/consumer evidence plus a reachable outcome; other statuses require a matching finding and action. Trace UI outcomes backward from render/action, and service/data outcomes through caller, boundary, side effect/persistence, and reload/reconciliation. Do not repeat requirement prose or evidence already cited elsewhere in the report.

Handoff

• Standalone: return verdict, reviewer runtime/model, fallback reason, report path, and only numbered CRITICAL/HIGH findings. Blockers recommend spectre-fix; otherwise recommend one of spectre-prove, spectre-test for a concrete gap, or spectre-clean only when proof is explicitly deferred.
• --orchestrated: return verdict, delivery counts, CRITICAL/HIGH evidence chains and identities, reviewer metadata, and report path; do not pause or suggest another command.

Escalate-If

• Scope remains ambiguous after artifacts and git state are inspected: ask what to review before dispatch.
• A proposed finding changes requirements: label Scope Change Required and exclude it from the blocking verdict.

Next step: follow the verdict-specific handoff.