jdsnack-agent-os-project: Skill for Claude Code

.agents/skills/review-loop/SKILL.md

review-loop is a skill for Claude Code, Codex from silverThunder09/jdsnack-agent-os-project. It costs 107 tokens per session (1,662 once invoked), scanned A, original, MIT.

A handoff loop in which Claude reviews a Codex coding change and decides whether it meets the project's requirements. It runs build, test, lint, and change-size checks before assigning a score and requesting revisions when needed.

In plain words
What is it for?
Checking pull requests, validating backend and frontend changes, enforcing a diff-size limit, scoring reviews, and sending unresolved issues back to Codex.
Why use it?
It creates a repeatable review and correction process before a change is merged.

Skill for Claude CodeCodex

Written for no agent in particular: nothing here depends on one. Also seen: installed under .agents/ (shared by several agents); mentions Codex.

This is silverThunder09/jdsnack-agent-os-project's own configuration. It tells Claude Code and Codex how to work on jdsnack-agent-os-project itself, so it is not a mod to install elsewhere. Copy it as a starting point and replace the rules that are about this project. Everything jdsnack-agent-os-project configures →

Reuse

Borrowing it

Nothing to install: this file belongs to silverThunder09/jdsnack-agent-os-project. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.

Copy the file
curl -O https://raw.githubusercontent.com/silverThunder09/jdsnack-agent-os-project/main/.agents/skills/review-loop/SKILL.md
Clone the repo
git clone --depth 1 https://github.com/silverThunder09/jdsnack-agent-os-project

Made for: Claude Code, Codex.

Wrote 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.

agentmods badge for review-loop

README.md
[![agentmods](https://agentmods.dev/badge/skills/silverthunder09/jdsnack-agent-os-project/review-loop/github.svg)](https://agentmods.dev/skills/silverthunder09/jdsnack-agent-os-project/review-loop)
Your own site
<a href="https://agentmods.dev/skills/silverthunder09/jdsnack-agent-os-project/review-loop"><img src="https://agentmods.dev/badge/skills/silverthunder09/jdsnack-agent-os-project/review-loop/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

agentmods 80×15 button for review-loop

Your own site · 80×15
<a href="https://agentmods.dev/skills/silverthunder09/jdsnack-agent-os-project/review-loop"><img src="https://agentmods.dev/badge/skills/silverthunder09/jdsnack-agent-os-project/review-loop.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 107 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 1,662 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.
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 5.1 $0.00107 $0.01662
Opus 5.5 $0.00043 $0.00665
Sonnet 5.5 $0.00021 $0.00332
Haiku 4.5 $0.00011 $0.00166

Measured 19d ago against content hash e6711f6a3e94, method: parsed. Prices are Anthropic first-party input rates as of 2026-10-07, from the pricing page.

Security

Grade A, and why

review-loop 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 19d 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/skills/review-loop/SKILL.md · 74 lines

How it starts

The opening of the file, as written. The whole thing — 74 lines — stays where its author put it; the contents beside it link to each section on GitHub.

review-loop (설정된 reviewer = 게이트키퍼)

역할 분담: Codex는 구현·테스트와 현재 기본 PR 리뷰 및 Spec 기획·문서 계획, Claude는 명시적으로 선택된 리뷰 backend 또는 사용자 승인 폴백을 담당합니다. reviewer 선택은 scripts/review-policy.json의 primaryReviewer를 따릅니다. 현재 primary reviewer는 Codex이므로 Claude 호출 없이 바로 읽기 전용 리뷰를 실행합니다. primary가 Claude이고 Claude가 unavailable이거나 구조화 결과를 반환하지 못하면 Codex read-only reviewer로 전환합니다. 유효한 Claude 판정은 Codex 결과로 덮어쓰지 않습니다. 어느 reviewer든 소스 코드를 수정하지 않습니다. 규칙 정본은 .agent-os/이며 여기서 재서술하지 않습니다.

0. 전제

  • Codex가 에이전트 코딩을 끝내고 작업 브랜치에 커밋·푸시한 상태여야 함.
  • 위험도·PR 흐름은 .agent-os/operations/pr-automation-loop.md를 따름.

1. 결정론 게이트 (LLM 호출 전 — 실패 시 리뷰 토큰 0)

Claude가 검증 목적으로 실행. 하나라도 실패하면 리뷰를 시작하지 않고 2번 형식의 변경요청을 만들어 Codex에 넘김:

  1. diff 줄수 한도 — 변경 1,000줄 초과면 PR 분할을 Codex에 요구하고 중단. git diff --stat origin/main...HEAD 합계(또는 gh pr diff <N> | wc -l) 확인.
  2. PR 계약 — bash scripts/pr-contract-test.sh <N>
  3. backend 변경 시 — cd backend && ./gradlew test
  4. frontend 변경 시 — cd frontend && npm run lint && npm test
  5. 모두 통과해야 2단계로 진행. (게이트용 재실행은 Claude가 수행, 코드 수정은 하지 않음)

2. 리뷰 핸드오프 루프 (최대 3회)

attempt = 1로 시작. 각 회차:

  1. diff 추출 — 로컬이면 git diff origin/main...HEAD, PR이면 gh pr diff <N>.
  2. code-reviewer 서브에이전트 호출 — 입력은 오직:
    • 위 diff
    • 대상 spec의 acceptance-criteria.md / test-scenarios.md 경로
    • attempt > 1이면 직전 회차 findings 중 미해결분(델타)만
    • 전체 레포·.agent-os 전체·대화 전사를 넣지 말 것.
  3. 점수 분기:
    • score ≥ 4 → 루프 종료(통과). 3단계로.
    • score < 4 → Claude는 수정하지 않는다. code-reviewer의 findings를 그대로 Codex용 변경요청으로 정리해 출력:
      → Codex 변경요청 (attempt N)
      score: N/5
      수정 항목(blocker/major 우선):
      - 파일:라인 — 무엇을 어떻게 고쳐야 하는가
      
      이후 Codex가 같은 브랜치에서 수정·커밋·푸시한다. 푸시 완료를 확인하면 attempt += 1 후 1번(결정론 게이트 포함)부터 재리뷰. 재리뷰는 델타만 평가.
  4. 시도 소진 — attempt == 3인데도 score < 4면 루프 중단. 자동 merge 금지:
    • 마지막 점수·미해결 findings를 사용자에게 에스컬레이션.
    • pr-automation-loop.md 기준 실패 Issue 생성 여부 판단.

3. 리뷰 결과 제출과 승인 job

리뷰 job은 구조화된 판정을 만들고, 승인 여부를 정해진 게이트에 맡깁니다. review job 자체에서는 PASS를 승인하거나 머지를 요청하지 않습니다.

Read the full file on GitHub · 74 lines

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. 19d ago Changed e6711f6a3e94
  2. 25d ago Changed fca3544b69a1
  3. 26d ago Changed · -10 tokens per session d71bafedc8f6
  4. 1mo ago Changed · -6 lines c2557b5af483
  5. 1mo ago Changed · +19 lines · +1 tokens per session f43429e3402b
  6. 1mo ago Changed · +3 tokens per session 9e965361baf0
  7. 2mo ago First seen · 61 lines · 113 tokens per session scan A 760d93396c1d

Subscribe to this mod's changes

review-loop is a skill published in the GitHub repository silverThunder09/jdsnack-agent-os-project (6 stars, last pushed 14d ago), licensed MIT. It adds 107 tokens to every session and 1,662 once invoked, about $0.0004 per session on Opus 5.5. 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.