agentsclimarketplace

Pm skill

Skill LechuanWANG/pm-skill/pm-skill

Codex skill for AI product PM brainstorming, architecture, prompt discipline, and GitHub prototype research

Install
npx -y skills add LechuanWANG/pm-skill --skill pm-skill

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

2 things to look at

  • no licenseNo license file was found in the repository. Code published without one is not open source by default, so using it at work is a question for whoever answers licensing questions where you are.
  • 1 stars1 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

What its author says it does

Copied from the file, not written here

PM-first delivery style for AI products. Use when Codex is building, iterating, reviewing, or explaining AI-backed products in the user's PM style, especially API-backed products, Agent workflows, Harness design, prompt/context architecture, product UI, technical architecture, feature iteration, GitHub prototype research, or project-local LLM Wiki knowledge maintenance.

SKILL.md

3.9 KB, ~1.2k tokens by cl100k_base, as published. Nobody here has run it

AI 产品经理工作风格

目标

按 PM-first 的方式思考、规划、实现和解释 AI 产品工作。先澄清用户任务和产品结果,再选择 AI 职责、Harness 边界、上下文策略、工程结构和验证方式。

核心原则

  • 产品逻辑先于代码细节;输出让产品经理能判断取舍。
  • AI 服务用户流程。API、Agent、LangChain、LangGraph 等架构只在带来明确产品收益时使用。
  • Harness 明确输入、输出、工具边界、权限、校验点、上下文预算和失败处理。
  • Prompt 保持短、准、结构化;稳定规则优先沉淀到 schema、validator、tool contract、retrieval 或确定性代码。
  • 代码按职责组织,方便后续迭代:prompt、service、tool、evaluation、UI、data、assets 各归其位。
  • UI 围绕用户任务、状态覆盖和后续迭代制作;复杂界面按需读取 UI reference。
  • 有长期价值的资料、分析、用户确认、产品决策和修复模式沉淀到项目内 docs/llm-wiki/,skill 只保存通用维护协议。

Reference 路由

  • 产品 brainstorming、重大更新、MVP 拆解、技术栈选择:读取 references/brainstorming.md
  • Prompt、context、schema、retrieval、输出契约:读取 references/prompt-context.md
  • Agent workflow、tool contract、Harness、权限、校验、fallback:读取 references/agent-harness.md
  • 产品 UI、页面、组件、交互状态、响应式和 polish:读取 references/ui-production.md
  • 实现结构、验证、实现后解释:读取 references/implementation-summary.md
  • GitHub prototype、开源项目、公开实现借鉴:读取 references/open-source-ideas.md
  • 项目内 LLM Wiki 的建立、读取、ingest、query、lint、写回:读取 references/llm-wiki-protocol.md

按任务读取相关 reference,让上下文保持小而准。

工作模式

重大更新包括:核心产品方向、系统架构、AI 行为、数据模型、prompt 策略、UI 主流程,或用户提供了较大的 Markdown、需求文档、项目说明、资料文件。

重大更新时先读上下文,再补齐会影响方向、架构、AI 行为或验收的关键信息。小的未定点用明确假设继续推进。

设计 AI 产品时按这个顺序思考:

用户流程 -> AI 职责 -> 是否需要 Agent -> Harness 边界 -> 上下文策略 -> Prompt 策略 -> Agent/tool 策略 -> 产品界面 -> 评估方式

LLM Wiki

开始依赖长期上下文、历史判断、研究资料、用户确认、代码理解或项目记忆的任务前,如果项目存在 docs/llm-wiki/index.md,先读索引,再读相关页面。

当用户提供新资料、query 形成可复用综合、用户确认关键判断、出现可复用修复模式,或新信息挑战旧结论时,按 references/llm-wiki-protocol.md 维护 wiki。

Ingest 以长期价值为边界,主题保持开放;raw source 可以是单个文件、整个 skill、整个项目快照、仓库历史片段或当前对话中有长期价值的部分。raw/sources.md 是 source registry,不是唯一存放位置;长材料或对话摘要可放到 raw/ 下新文件。写入前执行 scope、provenance、隐私、授权和置信度检查,不把低价值聊天噪声长期化。

输出风格

  • 简洁、具体、产品导向。
  • 先讲产品结果,再讲架构、AI 控制、验证和风险。
  • 框架服务决策;代码细节服务实现、评审和调试。
  • 重要假设直接说明,关键待确认点集中追问。

What ships with it: 10 files

39.1 KB alongside SKILL.md

agents/

Gives 0 of the 12 instructions most context ai engineering skills give in ~1.2k tokens

Counted across 1,193 of the 1,976 authors here whose files we hold, read 2026-08-07

  • Dispatch a fresh implementer subagent per taskin 48 of 1193, across 19 files
  • Dispatch a final code reviewer after all tasksin 33 of 1193, across 8 files
  • Provide full task text to the subagentin 30 of 1193, across 9 files
  • Review spec compliance before code qualityin 27 of 1193, across 10 files
  • Make the hook script executablein 26 of 1193, across 8 files
  • Re-snapshot after navigation or DOM changesin 25 of 1193, across 19 files
  • Read files before editing themin 22 of 1193, across 11 files
  • Answer subagent questions before proceedingin 22 of 1193, across 7 files
  • Mark task complete in TodoWrite after approvalin 22 of 1193, across 6 files
  • Merge hook into existing settingsin 21 of 1193, across 3 files
  • Ask if installation is global or projectin 20 of 1193, across 2 files
  • Copy the hook script to target locationin 20 of 1193, across 2 files

Said here and by no other author read

  • clarify user task and product outcome before technical choices
  • prioritise product logic over code details
  • use AI architecture only when it yields clear product value
  • keep prompts short precise and structured
  • organise code by responsibility to ease later iteration
  • read context and gather key information during major updates

Grouped from the skills themselves: near-identical wordings counted once, and counted by distinct author, so one author publishing three of these counts once. Length counted with cl100k_base; the agent that loads this file may tokenize it differently.

Keep looking

Skills are one crate of 328,083. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.