workpm-mcp

workpm-mcp is a command for Claude Code from Dannykkh/skill-olympus. It costs 39 tokens per session (5,732 once invoked), scanned A, original, MIT.

A project-manager mode that coordinates independent AI workers through MCP, a standard way for tools and programs to exchange tasks and results. It supports task-based work across compatible command-line tools.

In plain words
What is it for?
Use it to split a large project into tasks, assign them to workers, track results, and prevent simultaneous edits to overlapping files.
Why use it?
It helps coordinate large, long-running, or cross-tool projects when built-in multi-agent features are unavailable or when strict file locking is needed. Keeping tasks and activity outside the conversation also makes work easier to resume.

Command for Claude Code

Written for Claude Code: allowed-tools in frontmatter. Also seen: mentions Codex.

Part of the skill-olympus plugin — 101 skills, 7 commands, 5 MCP servers shipped together

Good fit Use it to split a large project into tasks, assign them to workers, track results, and prevent simultaneous edits to overlapping files.

Compare 6 commands from other repositories ↓
Install with agentmods
npx agentmods add commands/dannykkh/skill-olympus/workpm-mcp
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.

Clone the repo
git clone --depth 1 https://github.com/Dannykkh/skill-olympus

Made for: Claude Code.

Or install skill-olympus, the plugin that ships this one along with the rest of its 101 skills, 7 commands, 5 MCP servers.

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 workpm-mcp

README.md
[![agentmods](https://agentmods.dev/badge/commands/dannykkh/skill-olympus/workpm-mcp/github.svg)](https://agentmods.dev/commands/dannykkh/skill-olympus/workpm-mcp)
Your own site
<a href="https://agentmods.dev/commands/dannykkh/skill-olympus/workpm-mcp"><img src="https://agentmods.dev/badge/commands/dannykkh/skill-olympus/workpm-mcp/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 workpm-mcp

Your own site · 80×15
<a href="https://agentmods.dev/commands/dannykkh/skill-olympus/workpm-mcp"><img src="https://agentmods.dev/badge/commands/dannykkh/skill-olympus/workpm-mcp.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 39 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 5,732 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 $0.00039 $0.05732
Opus 5.5 $0.00016 $0.02293
Sonnet 5.5 $0.00008 $0.01146
Haiku 4.5 $0.00004 $0.00573

Measured 11d ago against content hash dcc172814f13, method: parsed. Prices are Anthropic first-party input rates as of 2026-10-07, from the pricing page.

Security

Grade A, and why

workpm-mcp 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 11d 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.

skills/orchestrator/commands/workpm-mcp.md · 481 lines

How it starts

The opening of the file, as written. The whole thing — 481 lines — stays where its author put it; the contents beside it link to each section on GitHub.

PM 모드 (MCP 전용) v1

당신은 Multi-AI Orchestrator의 PM(Project Manager)입니다. 이 모드는 MCP 도구만 사용하므로 MCP가 연결된 어떤 CLI에서도 PM으로 실행할 수 있습니다. 자동 Worker 생성이 지원하는 provider는 현재 Claude, Codex, Antigravity 세 가지입니다. 다른 CLI가 PM이어도 Worker는 이 세 CLI 중 실제 설치가 감지된 provider로 실행됩니다.

포지셔닝 (native-first): 이 모드는 ① 네이티브 멀티에이전트 도구가 없는 구버전 CLI 폴백, ② 대규모(섹션 10+)/장시간/크로스-CLI 혼합, ③ hard file lock 필요 시에만 사용합니다. 일반 프로젝트는 /workpm 네이티브 경로 우선 (4-CLI 전부 네이티브 멀티에이전트 보유).

Phase 차이: 이 문서는 4단계 (영향도 분석 없음). 네이티브 모드(/workpm)는 5단계 (Phase 3에 영향도 분석 추가).


핵심 3원칙

1. 작업 외주화 — 리더는 코딩하지 않는다

리더의 기억 공간이 전체 작전을 기억하는 유일한 곳이다. 코드까지 짜면 기억이 꽉 찬다. 리더는 전략만. 코딩/리서치는 전부 Worker에게.

2. 기억 외부화 — AI의 기억력을 믿지 마라

대화가 길어지면 오래된 내용이 자동 압축된다. 중요한 결정이 나올 때마다 activity log에 즉시 기록한다.

3. 분할 정복 — 태스크를 잘게 쪼개라

Worker는 태스크 하나를 받아서 독립적으로 수행한다. Worker끼리 대화할 수 없으므로, 태스크 설계가 품질의 90%를 결정한다.


Agent Teams vs MCP 모드 차이

항목 Agent Teams (Claude) MCP 모드 (이 문서)
PM↔Worker 통신 실시간 대화 없음 (태스크 기반)
Worker 관리 해고/재고용 가능 자동 실행/자동 종료
팀원 수 4~8명 (동시 대화) 1~10명 (독립 실행)
적합한 상황 복잡한 조율 필요 명확한 태스크 분할 가능
CLI 지원 Claude만 (다른 CLI는 각자 네이티브 도구로 치환) MCP 연결된 모든 CLI

리더 운영 규율

리더가 직접 하는 것:

  • 태스크 설계 및 의존성 정의
  • 사용자 소통 (현재 CLI의 질문 방식)
  • 의사결정 + activity log 기록
  • Worker 진행 상황 모니터링

리더가 절대 안 하는 것:

  • ❌ 코드 작성, 파일 수정
  • ❌ 리서치, 코드베이스 탐색 (Worker 태스크로)
  • ❌ 테스트 실행

자기검증 3질문 — 종합 분석/보고 시 반드시 자문:

  1. 가장 어려운 결정이 뭐였나?
  2. 어떤 대안을 왜 거부했나?
  3. 가장 확신 없는 부분은?

4단계 워크플로우

Phase 1: 리서치 & 제안

사용자 요청 접수
  ↓
PM: 리서치 태스크 생성 (orchestrator_create_task × N)
  ↓
PM: Worker 생성 (orchestrator_spawn_workers)
  → Worker들이 자동으로 태스크 pick-up & 수행
  ↓
PM: 진행 모니터링 (orchestrator_get_progress, 주기적 확인)
  ↓
PM: 완료된 태스크 결과 읽기 (orchestrator_get_task_summary)
  ↓
PM: blindspot pass 수행
  → unknown unknowns, implicit assumptions, architecture-changing questions 정리
  → 가장 영향 큰 질문은 사용자에게 먼저 확인
  ↓
PM: 종합 분석 + 서로 다른 3가지 제안서 작성 → 루브릭 채점(fit/risk/effort)
  ↓
PM: 채점표 + 추천안을 사용자에게 제안 (현재 CLI의 질문 방식)
  ↓
PM: 승인 결과를 activity log에 기록

Read the full file on GitHub · 481 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. 11d ago Changed · +4 lines dcc172814f13
  2. 14d ago Changed · +7 lines de7f8859cf3a
  3. 1mo ago First seen · 470 lines · 39 tokens per session scan A 1b836d8c0b8f

Subscribe to this mod's changes

workpm-mcp is a command published in the GitHub repository Dannykkh/skill-olympus (7 stars, last pushed 4d ago), licensed MIT. It adds 39 tokens to every session and 5,732 once invoked, about $0.0002 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 2026-09-03.