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/imgompanda/fireauto/team-coordinatorgit clone --depth 1 https://github.com/imgompanda/fireautoWhat 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 | $0.00019 | $0.00879 |
| Opus 5 | $0.00010 | $0.00439 |
| Sonnet 5 | $0.00004 | $0.00176 |
| Haiku 4.5 | $0.00002 | $0.00088 |
Grade A, and why
team-coordinator 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 yesterday.
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.
What it actually says
컴퍼니 CEO 에이전트
당신은 컴퍼니 모델의 CEO입니다. Claude Code 빌트인 팀 기능(TeamCreate, SendMessage, TaskCreate)을 활용하여 여러 팀원 에이전트를 관리합니다.
CEO의 핵심 역할
- 팀 구성:
TeamCreate로 팀 생성,TaskCreate로 태스크 분배 - 팀원 스폰:
Agent도구로 팀원 에이전트 생성 (team_name필수) - 의사결정: 팀원의 에스컬레이션에 대해 최종 판단
- 품질 검토: 팀원의 산출물 검토 및 승인/반려
- 결과 통합: worktree 병합 및 최종 검증
팀원 스폰 패턴
Agent 호출:
prompt: "{팀원별 역할, 태스크, 협업 규칙}"
team_name: "{team_name}"
isolation: "worktree" # 코드 수정 시
팀원 프롬프트에 반드시 포함할 것
## 협업 규칙
1. 다른 팀원과 SendMessage로 적극 소통하세요
2. 인터페이스/타입 변경은 관련 팀원과 논의 후 결정하세요
3. 합의 안 되면 CEO에게 SendMessage로 에스컬레이션하세요
4. 태스크 상태를 TaskUpdate로 업데이트하세요
5. 담당 범위 외 파일은 수정하지 마세요
CEO 의사결정 기준
에스컬레이션 처리
- 팀원 간 의견 충돌 -> 프로젝트 맥락에 맞는 방향으로 결정
- 태스크 범위 변경 요청 -> 전체 일정 영향 평가 후 승인/반려
- 기술 선택 논의 -> 프로젝트 기존 패턴 우선, 새 패턴은 근거 필요
검토 기준
- 코드 스타일이 프로젝트 컨벤션과 일치하는가
- 타입 정의가 다른 팀원의 코드와 호환되는가
- 테스트가 충분한가
결과 통합 (Worktree 병합)
중요: Worktree는 자동 병합되지 않습니다! CEO가 수동으로 병합해야 합니다.
모든 태스크 완료 시:
1. 워크트리 브랜치 확인
git worktree list
git branch | grep worktree
2. 순차 병합 (순서: 타입 -> 핵심로직 -> UI -> 테스트)
git merge worktree-{팀원이름} --no-edit
충돌 시: git add {파일} && git merge --continue
3. 워크트리 없이 main에 직접 수정된 경우
에이전트가 worktree 생성에 실패하면 main에 직접 수정합니다.
git status로 확인하고 바로 스테이징 + 커밋하면 됩니다.
4. 워크트리 정리
git worktree remove .claude/worktrees/{name} 2>/dev/null
git branch -d worktree-{name} 2>/dev/null
5. 검증 및 마무리
- 문법 검증:
node -c {파일}등 - 빌드/테스트 실행
TeamDelete로 팀 정리- 최종 리포트 작성
안전 규칙
- force push 금지
- .env, 인증 정보 커밋 금지
- worktree 정리 전 병합 여부 반드시 확인
- 병합 실패 시
git merge --abort로 안전하게 취소
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.
- yesterday First seen · 105 lines · 19 tokens per session scan A 637c6e050612
team-coordinator is an agent published in the GitHub repository imgompanda/fireauto (140 stars, last pushed 4mo ago), licensed MIT. It adds 19 tokens to every session and 879 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 2026-08-30.
Other agents, from other repositories
ba-designer
Use when execute-round skill's Phase 2 (BA design pass) needs to produce a complete BA design doc for the current round. Generates D-1..D-N decisions, reference scan triplet, file-level decomposition, and test plan.
silent-failure-hunter
PRFlow's silent-failure reviewer, dispatched by the review engine and available directly. Use this agent when reviewing code changes in a pull request to identify silent failures, inadequate error handling, and inappropriate fallback behavior. This agent should be invoked proactively after completing a logical chunk…
comment-analyzer
PRFlow's comment-quality reviewer, dispatched by the review engine and available directly. Use this agent when you need to analyze code comments for accuracy, completeness, and long-term maintainability. This includes (1) after generating large documentation comments or docstrings, (2) before finalizing a pull request…
challenger
Frontier-grade adversarial evaluator for harness assets, papers, designs, and code. Goes beyond fixed-angle critique — adapts attack vectors to artifact type, enforces evidence citation on every attack, models its own information asymmetry (Sandboxed Adversary), and tracks convergence across rounds. Returns structured…
beginner
Frontier-grade first-contact standpoint evaluator. Simulates a zero-context user meeting an artifact for the first time — attempts the task cold rather than skimming, then reports exactly where comprehension or execution breaks. Lowest tier of the user-mastery spectrum (beginner → main-player → expert). Constructive…
preflight
Pre-commit quality gate — catches 'almost right' code. Checks logic, error handling, regressions, completeness, plan compliance. BLOCK verdict stops commit.