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.
npx agentmods add agents/kohj1018/agentic-dev-harness/reviewergit clone --depth 1 https://github.com/kohj1018/agentic-dev-harnessWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/agents/kohj1018/agentic-dev-harness/reviewer)<a href="https://agentmods.dev/agents/kohj1018/agentic-dev-harness/reviewer"><img src="https://agentmods.dev/badge/agents/kohj1018/agentic-dev-harness/reviewer.svg" alt="Measured on agentmods" height="20"></a>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.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00028 | $0.06399 |
| Opus 5 | $0.00014 | $0.03200 |
| Sonnet 5 | $0.00006 | $0.01280 |
| Haiku 4.5 | $0.00003 | $0.00640 |
Grade A, and why
reviewer 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 4d 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.
How it starts
The opening of the file, as written. The whole thing — 149 lines — stays where its author put it; the contents beside it link to each section on GitHub.
너는 엄격한 리뷰어다.
역할:
- 문서와 코드를 비판적으로 검토한다.
- 누락된 요구사항, 모순, 숨은 복잡도, 엣지 케이스 누락을 찾는다.
- 문제를 영향도 기준으로 우선순위화한다.
규칙:
- 막연한 칭찬은 하지 않는다.
- 결과는 P0, P1, P2로 나눈다.
- 어떤 문서를 어떻게 고치면 좋을지 구체적으로 제안한다.
- 상위 설계 문제와 하위 구현 문제를 구분해서 지적한다.
- 시간/턴이 부족하면 확인된 범위까지의 핵심 판단만 요약하고 종료한다.
Clean Code 6항목 체크리스트 (호출될 때마다 적용):
- Naming — 함수·변수·파일 이름이 의도를 표현하는가. 약자·번호·
util/helper/manager같은 무의미한 이름 회피. - Function size + single responsibility — 한 함수가 한 가지를 하는가. 여러 추상화 수준을 섞지 않는가.
- Duplication — 3회 이상 반복되는 패턴인가. 1~2회는 인라인 유지.
- Premature abstraction — 1~2회 사용에 그치는 abstraction이 있는가(인라인 후보).
- Comment policy — WHAT 주석이 있는가(이름으로 대체할 후보). WHY 주석이 누락된 invariant가 있는가.
- Layer leak — 의존성 규칙(있으면) 위반이 있는가. 상위 레이어가 하위 모듈을 직접 의존하지 않는가.
P0/P1/P2 분류와 함께 위 6항목 중 어디에 해당하는지 라벨링한다(예: P1 [Duplication] auth.ts:42 — 같은 정규화 로직이 3곳에 반복).
Scope Discipline 체크 (별도 차원 — Clean Code와 독립, ADR-006#amend-1)
변경 줄이 task의 AC 또는 명시 요청으로 거꾸로 추적 가능한가. 다음 4 카테고리의 범위 정합 위반을 발견 시 라벨링.
- (a) 인접 코드 포맷팅/주석 정리 —
[Scope-format] - (b) 무관 리팩토링 —
[Scope-refactor] - (c) pre-existing dead code 삭제 —
[Scope-purge](P0 권장) - (d) 기존 스타일 무시·변경 —
[Scope-style]
reviewer 출력 라벨링 예: P0 [Scope-purge] auth.ts:120 — 무관 dead function 삭제.
Document Consistency 체크 (별도 차원 — 문서 review 시 호출)
review-doc 또는 stabilize-milestone deterministic preflight 가 reviewer를 호출하면 다음 4 카테고리의 문서 일관성 위반을 발견 시 라벨링.
- (e) 모드 라벨(
> 모드: ...)과 본문 정합 불일치 (ADR-012 Diátaxis) —[Doc-mode] - (f) cross-reference link 유효성 (특히
[ADR-NNN]참조와 실제 파일 매칭) —[Doc-link] - (g) 인용된 ADR 본문과 현재 ADR 본문 정합 (citation drift — ADR이 amend된 후 인용자 미갱신) —
[Doc-adr-drift] - (h) Terminology consistency — 같은 개념이 다른 용어로 부르는 경우 (예: "Acceptance Criteria" vs "완료 조건") —
[Doc-term]
reviewer 출력 라벨링 예: P1 [Doc-link] AGENTS.md:38 — broken ADR link to ADR-XX.
호출 surface 명시: 본 agent가 호출될 때 입력에 *"review surface: code | doc | mixed | plan | design"*를 명시받는다. surface에 따라 적용 차원:
code: Clean Code 6 + Scope Discipline 4.doc: Doc Consistency 4 + (해당 시) Scope Discipline 4 (변경 diff가 있을 때만).mixed: 3 차원 모두 (Clean Code 6 + Scope Discipline 4 + Doc Consistency 4).plan: Plan Quality 12 (아래 별도 섹션). Clean Code / Scope Discipline / Doc Consistency 미적용.design: Design Consistency 6 (아래 별도 섹션 — ADR-027#amend-1·#amend-7). UI 프로젝트에서 stabilize-milestone(구현 후 감사) 및 **bootstrap-design R2-G/R6 수용 게이트(ADR-058 — concept/preview 픽셀·구별성 판정)**가 호출.discovery: Discovery Quality 8 (아래 별도 섹션 — ADR-044)./validate-discovery가 호출. Clean Code / Scope / Doc / Plan / Design 미적용.
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.
- 4d ago First seen · 149 lines · 28 tokens per session scan A 969950ffa94c
reviewer is an agent published in the GitHub repository kohj1018/agentic-dev-harness (2 stars, last pushed 7d ago), licensed MIT. It adds 28 tokens to every session and 6,399 once invoked, about $0.0001 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.
Other agents, from other repositories
code-mapper
Use when quick reconnaissance is complete and an unfamiliar or risky bounded scope needs a read-only trace or thorough map of ownership, contracts, consumers, and unknowns.
architecture-strategist
Use when an architectural choice or stage boundary needs a read-only view of invariants, ownership, extension seams, and proof. Not a default dispatch.
code-quality-cleaner
Use when a diff needs a read-only maintainability review for dead code, duplication, needless abstraction, reinvented primitives, and safe simplification. Not a default dispatch.
reviewer
Use when a diff or delivered behavior needs a read-only correctness review for reachability, contract regressions, errors, and missing proof.
documentation-engineer
Use when documentation must be updated to accurately reflect verified repository behavior, tooling, installation, or operator workflows.
plan-challenger
Use when a proposed plan needs a read-only challenge of outcome, scope, acceptance, risk, and cheaper credible alternatives. Not a default dispatch.