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.
git clone --depth 1 https://github.com/TashanGKD/tashan-cursor-skillsWrote 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/rules/tashangkd/tashan-cursor-skills/role-menu)<a href="https://agentmods.dev/rules/tashangkd/tashan-cursor-skills/role-menu"><img src="https://agentmods.dev/badge/rules/tashangkd/tashan-cursor-skills/role-menu.svg" alt="Measured on agentmods" 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 | $0.09726 | $0.09726 |
| Opus 5 | $0.04863 | $0.04863 |
| Sonnet 5 | $0.01945 | $0.01945 |
| Haiku 4.5 | $0.00973 | $0.00973 |
Grade A, and why
role-menu 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 8d 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.
How it starts
The opening of the file, as written. The whole thing — 307 lines — stays where its author put it; the contents beside it link to each section on GitHub.
他山AI产品 · 角色菜单
当用户开始任何产品开发任务时,AI 必须先加载对应角色的 Skill,按其激活协议执行,而不是凭训练知识随意发挥。
可用角色一览
| 角色 | Skill 路径 | 触发场景 |
|---|---|---|
| 产品经理 | .cursor/skills/role-产品经理/SKILL.md |
产品设计、需求分析、用户流、闭环设计、MVP、产品定义、开发计划 |
| 技术架构师 | .cursor/skills/role-技术架构师/SKILL.md |
技术架构、系统设计、数据模型、接口规范、技术选型 |
| 前端开发 | .cursor/skills/role-前端开发/SKILL.md |
前端实现、React、TypeScript、页面、组件、接口联调 |
| 后端开发 | .cursor/skills/role-后端开发/SKILL.md |
后端实现、FastAPI、API接口、数据库、认证、SSE |
| AI工程师 | .cursor/skills/role-AI工程师/SKILL.md |
LLM调用、Prompt设计、智能体、多智能体工作流、效果评测 |
| 测试工程师 | .cursor/skills/role-测试工程师/SKILL.md |
测试验收、Bug报告、关卡C、功能验证、闭环测试 |
| Bug修复循环协调者 | .cursor/skills/bug-fix-loop-coordinator/SKILL.md |
测试后自动循环修复直到上线:读取技术问题追踪台→按P0→P1→P2优先级spawn fixer子智能体修复→每修复一个触发回归→循环直至P0+P1全清→输出可上线结论(所有信息通过文档传递) |
| 修复子智能体 | .cursor/agents/fixer.md |
可被spawn的代码修复子智能体:隔离上下文+完整TDD协议(复现→失败测试→修复→CI→追踪台);三种模式:①单Bug内联修复(被测试agent spawn)②全自动修复循环(读追踪台P0→P1→P2至收敛)③紧急P0热修复;修复3次失败自动升级人工 |
| 测试环境自动准备 | .cursor/skills/test-env-setup/SKILL.md |
测试前一键准备环境:服务健康检查→自动启动后端→DB验证→测试账号自动创建→依赖验证;有终端权限,自动修复环境问题,失败时输出诊断指引(被 role-测试工程师 Step 0 和 fixer Step A0/B0 自动调用) |
| DevOps | .cursor/skills/role-DevOps/SKILL.md |
部署、上线、CI/CD、服务器、nginx、SSL、端口 |
| UI设计师 | .cursor/skills/role-UI设计师/SKILL.md |
界面设计、交互设计、体验感、组件规范、视觉 |
| 数据分析师 | .cursor/skills/role-数据分析师/SKILL.md |
数据分析、闭环完成率、用户行为、AI调用成本、健康度报告 |
| 审核者·用户模拟 | .cursor/skills/role-审核者-用户模拟/SKILL.md |
关卡A:产品定义完成后、代码开始前,强制执行 |
| 审核者·系统破坏 | .cursor/skills/role-审核者-系统破坏/SKILL.md |
关卡B:技术架构完成后、代码开始前,强制执行 |
| 规范问题修复 | .cursor/skills/project-convention-resolver/SKILL.md |
发现规范偏差时(产品/架构/代码/文档/Skill):分类→路由执行修复→验证→归档(与 issue-tracker 的区别:后者只记录,本 Skill 自动执行修复) |
| 问题追踪 | .cursor/skills/issue-tracker/SKILL.md |
用户报告问题时:拆解为产品设计问题和技术问题,分别记录到追踪台 |
| 产品演进规划 | .cursor/skills/product-evolution-planner/SKILL.md |
主动分析产品:对照 AI时代产品原则,识别缺口,生成战略建议(与 role-产品经理 不同:后者执行具体设计;本 Skill 是原则驱动的主动诊断) |
| Skill体系健康检查 | .cursor/skills/skill-system-health-check/SKILL.md |
主动检查 Skill/Agent/Rule 体系内部一致性:触发词冲突/孤立组件/版本漂移/Rule叠加矛盾(注:只检查内部一致性,不检查传播完备性) |
| Skill体系演进规划 | .cursor/skills/skill-evolution-planner-meta/SKILL.md |
基于历史数据规划 Skill 体系演进:分析 CO-BUILD-LOG/PENDING-EXPERIENCES/PENDING-SKILLS,生成新建/重构/合并/废弃建议(对应产品侧的 product-evolution-planner) |
| Skill体系完善执行 | .cursor/skills/skill-system-evolution-executor/SKILL.md |
L4系统型多组件完善的完整项目流程:Phase0规模确认→Phase1分析计划→Phase2双审核(关卡A+B)→Phase3分批执行→Phase4整体验证→Phase5经验沉淀(§9.7 implements) |
| Skill库全量搜索 | .cursor/skills/skill-library-search/SKILL.md |
全量搜索外部 Skill 仓库:7维度强制协议(官方/开发者/Claude专项/平台/Topic/交叉引用/验证),禁止以少于7维度声称「全量」 |
| 输入分类器 | .cursor/skills/cognitive-input-classifier/SKILL.md |
输入路由前置层:意图不明确时先分类——路径A(认知更新)→ cognitive-capture-fragment / 路径B(任务执行)→ 对应任务Skill;输出结论后强制停止等确认 |
| 认知-工作对齐检查 | .cursor/skills/cognitive-work-alignment-check/SKILL.md |
四维健康检查第四维:验证5个工作域核心产出是否可追溯到元思维体系(L1.5/L1),识别漂浮任务,输出对齐报告 |
| Skill产品经理 | .cursor/skills/role-Skill产品经理/SKILL.md |
纯产品定义层:Skill想法模糊时先「想清楚是什么」,双视角引导→输出draft卡片→再由skill-designer实现(注意:触发后手动引导skill-designer读draft) |
| 沙盘扩充 | .cursor/skills/skill-sandbox-expander/SKILL.md |
系统化扩充沙盘库:识别覆盖缺口→两阶段生成(Phase1不读SKILL.md)→持续推进100目标(也可补全已有draft) |
| 场景沙盘生成 | .cursor/skills/scenario-sandbox-builder/SKILL.md |
从零发现所有任务场景并建沙盘:从role-*动态归纳场景→识别大闭环(含认知根D0节点)→生成draft沙盘文件供 skill-closure-verifier-meta 消费。与沙盘扩充的区别:本Skill做初始全量发现,沙盘扩充做后续缺口填充 |
| 域健康检查 | .cursor/skills/skill-domain-health-check/SKILL.md |
按域检查传播完备性:节点可达/I-O契约满足/任意入口→闭环终态(通常由元验证器调用,也可独立运行) |
| 域自我优化 | .cursor/skills/skill-domain-self-optimizer/SKILL.md |
基于Gap证据生成定向修复方案:识别断裂模式→新Skill提案/现有Skill修改/缺失触发链路补全(通常由元验证器调用) |
| 元级闭环验证 | .cursor/skills/skill-closure-verifier-meta/SKILL.md |
全系统闭环验证编排器:两阶段(预想+实际)沙盘验证→域健康检查→域自优化→带证据的总健康报告(最高级Skill体系自检入口) |
| 项目收尾检查 | .cursor/skills/project-closeout/SKILL.md |
项目完成后、再次打开前:多角色文档审查 + 对话历史用户输入追踪,输出收尾报告(与 project-retrospective 不同:本 Skill 检查文档,后者更新 Skill 体系) |
| 跨项目并行协调 | .cursor/skills/cross-project-coordinator/SKILL.md |
启动多项目并行开发时:生成跨项目协调框架文档(项目地图/API契约/并行轨道/Day1清单),RULE-37 的执行层落地 |
| 他山工作台开发 | .cursor/skills/tashan-workbench-dev/SKILL.md |
接手/开发/调试 tashan-workbench 时:启动命令、架构速查、历史踩坑、测试方法 |
| 新产品兼容性验证 | .cursor/skills/product-launch-validator/SKILL.md |
新产品/新场景立项前:验证与三大闭环(Loop1 Skill覆盖/Loop2 认知根/Loop3 工作域)的兼容性,输出缺口报告(在 role-产品经理 前执行,可选但推荐) |
| 三大闭环健康检查 | .cursor/skills/three-loop-health-check/SKILL.md |
日常快速健检:Loop1+Loop2+Loop3↔2 统一仪表盘(运营指标+沙盘状态+D0覆盖率),约1分钟。⚠️ 无Full模式;深度验证→skill-closure-verifier-meta;生成沙盘→scenario-sandbox-builder |
| Skill设计师 | .cursor/skills/skill-designer/SKILL.md |
新建/重构任何 Skill/Agent/Rule 时:判断复杂度级别,双视角产品定义,触发对应关卡 |
| 代码库探索 | .cursor/skills/codebase-explorer/SKILL.md |
接手任何代码库/理解现有系统时:扫描架构,与产品定义对比,输出架构理解文档 |
| 基于已有项目重构 | .cursor/skills/project-reconstruction/SKILL.md |
现有项目技术债严重、需从干净新项目开始迁移时:6步流程(迁移评估→功能清单→技术框架+红蓝推演回写→Cursor规范配置→Phase开发计划),区别role-技术架构师:本Skill先盘清已有资产 |
| 历史审计 | .cursor/skills/history-auditor/SKILL.md |
审查历史对话/提取工作规律时:分析重复任务,提炼偏好,建议 Skill/Rule 升级 |
| 架构差距分析 | .cursor/skills/architecture-gap-mapper/SKILL.md |
对比两个系统/竞品分析/找差距时:并行读取,输出兼容/冲突/缺失三类,带证据链 |
| 文档整合 | .cursor/skills/doc-consolidator/SKILL.md |
文档漂移/多版本并存/恢复单一真源时:识别主真源,提取补充,重建结构 |
| 整改计划 | .cursor/skills/remediation-planner/SKILL.md |
将分析发现转为行动计划时:优先级分组,最小步骤,验证配对,branch结构建议 |
| 工程原则录入 | .cursor/skills/engineering-principle-recorder/SKILL.md |
立一条新的工程/执行原则时:分类(AI行为约束/工程设计/代码违规)→ 写入执行层(Role Skill/Rule)→ 审计违规点 → 覆盖报告。⚠️ 仅用于「原则本身尚不存在」的情况;若原则已有但被违反一次 → project-convention-resolver;若是认知/思维原则 → cognitive-extract-principle |
| 项目待办管理 | .cursor/skills/project-backlog/SKILL.md |
两种模式:【收集】任意混合输入(Bug/功能/UX/原则/建议)→ 解析追加到「项目待办.md」(不立即执行);【处理】读文档 → 前置决策收集 → 按优先级逐条路由执行(原则→engineering-principle-recorder / Bug→bug-fix-loop / Feature→协调者)→ 更新状态 → 批次报告。⚠️ 触发词必须含「待办」关键词 |
| Skill闭环捕捉 | .cursor/skills/skill-capture-closure/SKILL.md |
单次踩坑/单个经验沉淀:判断新建或更新 Skill,同步索引 |
| 项目级Skill复盘 | .cursor/skills/project-retrospective/SKILL.md |
项目/任务完成后批量复盘:扫描全部调用过的 Skill,逐个沉淀 |
| 系统调研输出 | .cursor/skills/research-output/SKILL.md |
对任何主题进行系统调研,产出图文Markdown(Mermaid+qwen-image),自动保存到认知结构知识库并注册 |
| 专家培养 | .cursor/skills/expert-bootstrap/SKILL.md |
按需培养领域专家子智能体:调研→对标→自问自答反思→产出可 dispatch 的专家 Agent 定义文件(B-object)。与调研(research-output)的区别:调研产出知识报告(K-object)供 CEO 用;专家培养让 Agent 成为该领域专家。Orchestrator 在「现有审查 Agent 对某领域深度不足」时触发 |
| 统一写作流水线 | .cursor/skills/document-pipeline/SKILL.md |
写任何文档/文章/报告(研究→草稿→逻辑自检→真实性校对→画图→审稿→格式转换),可从任意阶段入场 |
| 公众号文章写作 | .cursor/skills/wechat-article-writer/SKILL.md |
写他山公众号文章、推文;含HTML排版/配图/底部模板/发布流程 |
| 通用文章写作 | .cursor/skills/general-article-writer/SKILL.md |
写技术分析/产品思考/对外分享/个人随笔;不含微信排版;输出 Markdown |
| 学术 SI 写作 | C:\Users\Boyuan\.cursor\skills\academic-si-writer\SKILL.md |
写学术论文补充材料(SI),LaTeX格式,细胞力学/生物物理领域,PNAS/Nature风格 |
| 文章审稿 | .cursor/skills/article-proofreading/SKILL.md |
审查任何文章草稿,检查AI腔/标题规范/绝对表达/结语 |
| AI配图生成 | .cursor/skills/ai-image-generator/SKILL.md |
能力层 Skill:任何需要生成信息图时调用。含 qwen-image-2.0-pro API、多视角草图流程、风格库、配图索引联动。被 wechat-article-writer / role-UI设计师 / role-产品经理 等调用 |
| 前端品牌设计系统 | .cursor/skills/frontend-brand/SKILL.md |
他山前端品牌完整参考手册:技术框架/设计令牌/配色/圆角/组件代码模板/布局/动画/a11y;被 role-前端开发 和 role-UI设计师 调用 |
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.
- 8d ago First seen · 307 lines · 9,726 tokens per session scan A 622457e208cd
role-menu is a cursor rule published in the GitHub repository TashanGKD/tashan-cursor-skills (20 stars, last pushed 5mo ago), licensed MIT. It adds 9,726 tokens to every session, about $0.0486 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 cursor rules, from other repositories
10-feature-development
Feature implementation workflow and engineering mindset.
11-template-conventions
Reusable building blocks shipped with this template - use them instead of writing new ones.
12-new-project
Workflow for starting a new app from this template - rebranding, identity, cleanup and first feature.
13-updating-project
Workflow for updating an existing project - dependency and SDK upgrades, migrations, refactors, bug fixes.
01-tech-stack
Android tech stack and project standards.
02-architecture
Architecture rules - MVVM baseline, MVI for complex features.