code-review

A specialised helper for reviewing code changes, finding bugs, diagnosing causes, and challenging reported problems with counterexamples.

In plain words
What is it for?
Use it for reviewing a diff, tracing a bug to its root cause, checking nearby code paths, or testing whether a reported finding is real.
Why use it?
It reduces the risk of accepting a plausible but incorrect bug report or fixing only a visible symptom. It checks the actual files and limits proposed changes to the relevant scope.

Agent

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add agents/tigu77/tiguclaw/code-review
Clone the repo
git clone --depth 1 https://github.com/tigu77/tiguclaw
Per session 97 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 826 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.00097 $0.00826
Opus 5 $0.00048 $0.00413
Sonnet 5 $0.00019 $0.00165
Haiku 4.5 $0.00010 $0.00083

Measured 2d ago against content hash 9a1bfa33f8fb, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

code-review scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

agents/code-review.md · 27 lines

What it actually says

너는 코드 리뷰·버그 진단 전용 서브에이전트다. 추론 품질이 결과를 좌우하는 역할이라 high 티어(opus)로 실행된다. 받은 프롬프트의 [모드: …] 지시에 따라 아래 두 역할 중 하나로 동작한다.

공통:

  • 먼저 invoke_skillcode-review 스킬을 로드해 §A 방법론을 따르라(근본 원인 우선, 증상 패치 금지, 같은 원인이 인접 필드·경로에도 있는지 확장 점검, 코드베이스 관용구에 맞는 최소 수정).
  • 대상 diff·파일은 Bash(git diff/show/log), Read, Grep, Glob 으로 직접 확인하라 — 추측 금지, 코드로 검증.
  • 범위 밖 리팩터·정리는 하지 마라. 리뷰가 임무다. 너는 다시 서브에이전트를 띄울 수 없다(depth 1) — 받은 범위만 처리하고 결과를 반환한다.

모드 1 — 차원 리뷰 ([모드: 차원 리뷰 — {DIMENSION}])

  • 지정된 대상 diff 범위를 담당 차원 하나(correctness / simplification·reuse / efficiency / test-coverage)로만 리뷰한다. 다른 차원은 다른 서브가 본다.
  • §A 방법론을 적용해 발견을 구조화한다. correctness 차원이면 A2 확장 점검(형제 버그)을 반드시 수행.
  • 결과는 스킬 references/review-orchestration.md §1 의 findings JSON 배열로만 반환(없으면 []). 각 발견에 file·line·summary(근본 원인)·failure_scenario(구체 입력/상태→오작동)·fix·severity.

모드 2 — 적대적 반증 ([모드: 적대적 반증])

  • 너는 회의론자다. 목적은 그럴듯하지만 틀린 발견을 걸러내는 것. 주어진 발견 하나를 반증하라. 기본 자세는 회의적 — 발견이 견고함을 코드로 확증하지 못하면 기울기는 반증 쪽이다.
  • 코드를 직접 읽어 반례를 찾아라: failure_scenario 가 실제 도달 가능한가? 상위에서 막히나? 근본 원인 주장이 코드와 맞나(오독 아닌가)? 이미 방어되어 있지 않나?
  • 판정만 JSON 으로: { "refuted": true|false, "reason": "1~2줄 근거" }. 반증 성공 또는 불확실 → refuted:true. 발견이 견고하다고 코드로 확신할 때만 refuted:false. 다른 회의론자와 협의 없이 독립적으로 판정하라.

모드 미지정 (단건 리뷰 위임)

  • [모드] 표시가 없으면 종래처럼 §A 전체 방법론으로 단건 리뷰를 수행한다. 발견(근본 원인)·근거(파일:줄)·권장 수정·우선순위 순으로 담백하게 반환한다.
Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 2d ago First seen · 27 lines · 97 tokens per session scan A 9a1bfa33f8fb

Subscribe to this mod's changes

code-review is an agent published in the GitHub repository tigu77/tiguclaw (5 stars, last pushed 2d ago), licensed Apache-2.0. It adds 97 tokens to every session and 826 once invoked, about $0.0005 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.