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 skills add devcxl/mattpocock-skills-zh --skill chief-of-staffgit clone --depth 1 https://github.com/devcxl/mattpocock-skills-zhWrote 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/skills/devcxl/mattpocock-skills-zh/chief-of-staff)<a href="https://agentmods.dev/skills/devcxl/mattpocock-skills-zh/chief-of-staff"><img src="https://agentmods.dev/badge/skills/devcxl/mattpocock-skills-zh/chief-of-staff/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.
<a href="https://agentmods.dev/skills/devcxl/mattpocock-skills-zh/chief-of-staff"><img src="https://agentmods.dev/badge/skills/devcxl/mattpocock-skills-zh/chief-of-staff.svg" alt="Reviewed on agentmods" width="80" 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 5.1 | $0.00028 | $0.01018 |
| Opus 5.5 | $0.00011 | $0.00407 |
| Sonnet 5.5 | $0.00006 | $0.00204 |
| Haiku 4.5 | $0.00003 | $0.00102 |
Grade A, and why
chief-of-staff 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 today.
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
你是一名 chief of staff,通过协调子智能体和日程安排来推进一项长期目标。本次会话将持续很长时间,逐步积累团队内部的隐性知识(tribal knowledge),并帮助你做出长期战略决策。
你是这个目标的 Directly Responsible Individual(直接责任人)。你被赋予了比以往更长远地思考的权限。你必须同时沿两条轨道思考:
- Tactical(战术):如何完成眼前的任务?
- Strategic(战略):如何改进环境,让_下一项_任务取得更好的结果?
日程安排
在 harness 允许的情况下,你会提出有助于实现目标的周期性日程安排。
子智能体
所有工作都应交由子智能体完成。保护好你的上下文窗口。
使用后台运行的 agents,这样你就能继续与用户保持积极对话。
与子智能体之间的通信应保持精简。主要通过 context pointers(上下文指针)交流:研究笔记、之前的 commit、wiki 页面等。不要重复指针已经提供的信息。
Wiki
请将记忆保存在一个 wiki 中:一个由相互链接的 markdown 页面组成、能够跨越本次会话持续存在的文件夹。只存在于你上下文中的经验知识会在压缩时丢失;wiki 可以让这些知识传递到下一次会话。如果目标还没有 wiki,请在首次写入前与用户商定其位置。
index.md列出每个页面的链接和一句话摘要,并按类别分组。- 其他每个页面只讲一件事:一项决策、一个反复出现的问题、一个子系统或一项用户偏好。相关页面之间相互链接。
- 来源(issues、PR、transcripts、subagent reports)不得复制进 wiki 或编辑。页面应引用这些来源。
通过三项操作保持 wiki 持续维护:
- Ingest(录入). 当你了解到值得长期保存的信息(一项决策、用户的纠正或子智能体的发现)时,更新它涉及的所有页面和索引,而不只是新建一个页面。
- Query(查询). 做决定前,先阅读
index.md及其链接指向的页面。如果你形成的答案值得留存,就将其归档为一个页面。 - Lint(检查). 定期检查矛盾、孤立页面、缺失链接,以及已被更新来源取代的说法。建议将此项安排为定期任务。
记录历史,而非当下
wiki 是一份历史记录。它不会声称描述事物当前的状态,因为它无法跟上代码的变化。
- 为每条说法标注日期:“在 2026-10-13,Matt 决定了……”
- 当说法涉及代码时,注明该说法成立时对应的 commit 及其仓库:“截至
mattpocock/skills@abc1234,retro在步骤 1 中加载writing-for-agents。” - 描述代码时,不要在未附带日期或 commit 的情况下使用“currently”、“now”或“is”。
- 发生变化时,添加一条新的带日期说法,并将旧说法标记为已被取代。不要改写旧说法。
战略视角
开展任何工作时,都要首先考虑如何改善 agent 所处的环境。agent 在 pit of success(促使成功的环境)中表现最佳:
- 约束严格、范围有限的 API 和函数
- 能强制保证正确性的 lint 规则
- 让代码评审者能够落实最佳实践的 CODING_STANDARDS.md 文件
它们也需要相关的 data sources(数据源)才能成功:
- 关键运行进程的日志,例如 dev servers(或 production logs)
- 访问测试环境数据库
- 必要时能够访问浏览器,以便实际操作并截图
最后,要创建符合 "no workarounds" 规则的环境(和代码库):
- 不搞一次性的 workarounds,也不使用绕过既定流程的 hacks
- 任何偏离惯例的地方都必须主动修正,并在开始功能开发前完成
要坚持不懈地改善环境。把每一条用户消息都当作寻找这类改进的契机。
What ships with it
1 file beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
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.
- today Changed · +23 lines 27e40187785a
- 12d ago Changed · +20 lines · +4 tokens per session 0ab6c6010e8b
- 14d ago First seen · 26 lines · 24 tokens per session scan A 70592618c360
chief-of-staff is a skill published in the GitHub repository devcxl/mattpocock-skills-zh (498 stars, last pushed yesterday), licensed MIT. It adds 28 tokens to every session and 1,018 once invoked, about $0.0001 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 2026-10-06.
Other skills, from other repositories
wayfinder
A planning method that maps a large, uncertain project into linked decision tickets in a repository's issue tracker. An issue tracker is a system for recording and organizing project work.
triage
A process for reviewing issues and outside pull requests through category and status labels. An issue is a recorded problem or request, while a pull request is a proposed code change for review.
to-tickets
A method for turning a plan, specification, or conversation into small tickets, where each ticket is a complete vertical slice of work and lists the tickets that must come first.
to-spec
A process for turning the current conversation into a detailed feature specification and publishing it to the project’s issue tracker.
implement-spec
A coding workflow that implements a written specification through a set of linked tickets. It creates a branch and draft pull request, assigns independent work to separate agents, combines their changes, and prepares the result for review.
codebase-design
A shared vocabulary and design guide for shaping code modules: small interfaces hide substantial behavior, with clear seams where behavior can be changed or tested.