product-brief
Skill100 starsChanged 10 days agoMITView on GitHub
把一句需求或评分卡转成可验收的 PRD 与执行计划。用户要写产品需求、界定 MVP、拆里程碑(如「帮我把这个需求写成 PRD」)时调用;不打分、不编码、不替用户批准。
10 other projects use a version of this.
Adopted by
Most recently taken up by
- Job-Yang/jobyang-ai-skillsword for word, first seen Nov 6, 2026
- MarkTripoli/skillstheir own version, first seen Nov 1, 2026
- vindreshsingh/engineering-skillstheir own version, first seen Oct 20, 2026
- rterakedis/leanwheel-skillstheir own version, first seen Oct 17, 2026
- emotixco/claude-skills-foundertheir own version, first seen Oct 7, 2026
- lglucas/ai-dev-operating-systemtheir own version, first seen Sep 28, 2026
Used with
Projects with it also use decision-log and setup.
Other files people use alongside this one.
Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
nandanosqlSkilldecision-log
Document product decisions with context, alternatives considered, rationale, and revisit criteria. Creates institutional memory that prevents re-litigating settled decisions.
Also in 3 projects that use this
VercelSkillsetup
Set up Vercel CLI and project configuration. Use when the user says "set up Vercel", "configure Vercel", "link to Vercel", or "vercel init".
Also in 3 projects that use this
Other versions10 projects use a version of this.
Word for word · 1
- Job-Yang/jobyang-ai-skills★ 100100% same
Edited · 0
None found.
Wrote their own version · 9
- emotixco/claude-skills-founder★ 796same name
- davekilleen/Dex★ 509same name
- lglucas/ai-dev-operating-system★ 12same name
- MarkTripoli/skills★ 2same name
- rterakedis/leanwheel-skills★ 2same name
- tolgalogy/agentic-development-framework★ 2same name
- erik-kroon/skills★ 1same name
- philpalmieri/engineering-manager-skills★ 1same name
- vindreshsingh/engineering-skills★ 0same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository25 other files it gives its agents.
- ai-briefingSkill
- argument-readingSkill
- claim-checkSkill
- content-infographicSkill
- cuihuoSkill
- engineering-docsSkill
- feynman-explainerSkill
- haohao-evalSkill
- haohao-shuohuaSkill
- me-contextSkill
- minimum-experimentSkill
- opportunity-radarSkill
- opportunity-scorecardSkill
- outside-viewSkill
- png-compressSkill
- reverse-teardownSkill
- root-rethinkSkill
- sansi-erhouxingSkill
- skill-library-keeperSkill
- steelman-choiceSkill
- style-compassSkill
- talk-smithSkill
- tangshan-styleSkill
- vddSkill
- video-readerSkill
The file
Files
产品需求与计划
版本口径(2026-10-16 起):信封 schema 版本 0.2.0,由冻结的 scripts/artifact_io.py seal 出,代表信封结构;本 kit(技能包)版本 0.3.1,见 CHANGELOG.md。过去本文档自称"版本 0.2.1"的写法已废弃,不再与信封版本混用。先读包内 references/handoff.md。本技能只写产品产物;不自动编码、不自动发布,也不替用户批准计划。
关于包内 scripts/artifact_io.py 和 scripts/scoring-rules.json:它们是 opc-envelope 0.2.0 的 vendored 公共副本(见 scripts/vendor-manifest.json),作用只是给本节点的 prd/plan 产物做信封校验、打包和导入。本节点不执行评分,不读 scoring-rules.json 的权重做计算;需要评分时用 opportunity-scorecard 包。
输入与范围
接受直接描述、附件、显式文件路径或已批准的上游工件。主题明确即可开始,不强迫独立用户先跑雷达或评分。OPC 模式由当前 flow 明确指定,复用其输入、审批和输出目录;不重复询问已确认事项。
先提取目标人群、待解决问题、使用频率、关键场景、资源限制和已有决定。把事实、用户确认与假设分开。缺输入时先完成可独立推进的部分,只询问会改变实现范围的关键歧义。
执行
- 界定问题。 写清谁在什么场景遇到什么障碍、目前怎么解决、成功后如何变化。避免从功能清单倒推问题。
- 明确 MVP。 给每个功能稳定的 F 标识(F1、F2……编号必须连续,断号就地说明),记录触发条件、预期结果和可验证的验收条件。验收优先写"可注入/可回放的系统行为用例 + 频次上下限"(如"注入 20 个应抑制事件 0 触发"),用户行为指标次之;"日志留存完整率=100%"这类自证型内部指标不当验收,除非连着用户可感知后果。列出 out-of-scope,核心旅程只保留证明目标所必需的能力。
- 提出平台建议。 结合用户、使用频率和资源给推荐与替代方案,说明利弊。用户已决定则复用;不能擅自更换 Web、移动端、桌面端等形态。没有依据的收益、周期和流量数字标为待测目标。
- 整理 PRD。 使用
templates/prd.md的字段,不机械套未相关章节。数据模型保持产品层面(有哪些实体、各是什么),不写存储选型、接口契约、并发方案。"需要展开时交 engineering-docs"的可操作判据:出现下列任一情况,技术展开交该技能——① 要在两个技术方案间取舍(自建 vs 第三方、是否迁移旧系统),且结果影响工期量级;② 涉及系统间集成、数据迁移或合规约束;③ 计划里必须先做技术探查(Spike)才能估算。只是把需求翻译成页面字段和流程步骤的,留在本 PRD 内。 - 整理计划。 使用
templates/plan.md。写里程碑、依赖、估算依据与不确定区间、交付物、验收责任及回退范围。上游点名的弱项(如评分卡低维)必须按 plan.md"弱项验证四件套"落地:可测假设、证伪指标、止损动作、最小样本量(用样本数门槛,不编行业转化率)。设计和技术研究可以并行,不默认所有工作串行。 - 检查并交付。 每个 F 都有验收,计划引用的 F 存在。沿核心旅程检查会影响用户结果的输入规则:必填、格式、单位与精度、边界值、无效输入;已明确的写入验收,缺依据的列为待决或明确建议,不替用户拍板。只展开本任务相关项,局部改动不强加整套产品规划。未决问题要有解决方式。输出 PRD 与计划,均为待人审。完整计划不是开工授权,已得到的用户授权则明确记录。
如果需要外部信息才能选择方案,按需复用 outside-view;现有证据足够则直接工作。没有该技能也不能伪造行业共识或阻塞本地成文。
输出
prd-<topic>.md 与 plan-<topic>.md 使用同一 topic。两者 opc_node: prd,分别 artifact_kind: prd、artifact_kind: plan。直接输入的 upstream 为 null;计划可以引用本次 PRD。保留确认记录,不在机器自生成的头里写 approved。
正文完成后,用包内脚本封装与校验:
python3 scripts/artifact_io.py seal --input prd-body.md --node prd --kind prd --topic example-product --producer standalone --out prd-example-product.md
python3 scripts/artifact_io.py seal --input plan-body.md --node prd --kind plan --topic example-product --upstream prd-example-product.md --out plan-example-product.md
python3 scripts/artifact_io.py pack --input prd-example-product.md plan-example-product.md --out result.zip无 Python 时按 references/handoff.md 手工输出完整信封与正文,注明结构尚未机验。示例值不能冒充真实调研、工期或用户确认。
与相邻能力的边界
以下技能均为可选的后续能力,不是本技能的前置依赖。未安装时仍独立交付 PRD 和计划,展开到界面或工程实现之外的问题则标明交接范围。
- opportunity-scorecard 判断潜力,本技能定义交付范围;不重复打分。
- style-compass 负责界面结构、交互与视觉;本技能提供稳定需求和旅程,允许设计反馈修订。
- engineering-docs 负责工程取舍与技术方案;交接判据见执行第 4 步。方案改变工作量后更新计划版本及变更原因。
- 流程或用户决定下一步;本技能不强制生成新会话、调用下游或操作项目审批。
维护说明
本包自带 opc-envelope 0.2.0 公共文件副本(scripts/artifact_io.py、scripts/scoring-rules.json,见 scripts/vendor-manifest.json),本轮冻结不改内容。发布前跑 python3 scripts/check_vendored.py 自检;与其他流水线包并排时跑 python3 scripts/check_vendored.py --drift <各技能根> 确认无漂移,跑 opportunity-radar/scripts/check_handoff.py <三包根> 确认 handoff 示例 opc_node 无 drift。
每日定时由宿主 cron/云电脑拉起本技能;本技能只管产出 prd/plan 与信封契约,不负责调度、不内置 token、不自动发布。
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

