cuipengfei
- 26 skills
- 0 followers
- 1 week ago last updated
- ▌ Improve Skill · cuipengfei当用户要求改进已有 skill,或某个 skill 本次使用不顺、漏了指引、需要把有效经验补给下一个 agent 时使用。查 `~/.agents/.skill-lock.json`:未追踪的本地自制 skill 可就地改;已追踪的 skill 只在原 repo 改。分析后先提案,获用户批准才编辑。
- ▌ Github Repo Osint · cuipengfei对任意一个或一组 GitHub 仓库做情报研究时使用:产出分层简报——身份机制、硬指标、社区口碑、star 真实性档位、选型含义,每条结论带证据分级。触发于「调研这个仓库」「这个 repo 靠谱吗」「对比这几个仓库」「repo 口碑」「star 真不真」「开源选型尽调」等。
- ▌
- ▌
- ▌
- ▌ Fuck · cuipengfei当用户消息出现 "fuck"、"wtf"、"靠"、"操"、"shit"、"这都错了"、"你在干嘛" 等词时无条件加载。加载后回看相关上下文,判断是否指向 agent 刚才的理解、范围、假设、提问或动作;若明确指向第三方、外部事件、引用内容或单纯发泄,则不执行纠偏。
- ▌
- ▌
- ▌
- ▌ No Bs · cuipengfei当用户明确要求直接 verdict、说 "no-bs"、"no bullshit"、"be direct"、"stop being diplomatic"、"stop hedging"、"pick a side"、"fake fairness",或要求纠正假公平、无效免责声明、伪中立时使用。适用于明确不要 sycophancy 或外交腔的技术比较、诊断、决策、批评场景。
- ▌
- ▌
- ▌ Openspec Explore · cuipengfeiEnter explore mode - a thinking partner for exploring ideas, investigating problems, and clarifying requirements. Use when the user wants to think through something before or during a change.
- ▌ Openspec Propose · cuipengfeiPropose a new change with all artifacts generated in one step. Use when the user wants to quickly describe what they want to build and get a complete proposal with design, specs, and tasks ready for implementation.
- ▌ Multi Source Inquiry · cuipengfei当答案需要联网检索或独立多源验证,包括事实核查、技术对比、当前信息、冲突说法、重要建议,以及需要多轮深挖、迭代收敛的深度研究时使用。
- ▌ Exploring Generative AI · cuipengfei bundle使用 Martin Fowler 的 Exploring Generative AI 系列评估 coding assistant、agent 自主性、harness engineering(guides 与 sensors、feedback 与 feedforward、computational 与 inferential controls)、AI 辅助开发工作流、context engineering、本地 coding model 可行性或软件工程取舍时使用。
- ▌
- ▌ Openspec Apply Change · cuipengfeiImplement tasks from an OpenSpec change. Use when the user wants to start implementing, continue implementation, or work through tasks.
- ▌ Evolutionary Architecture · cuipengfei bundleKnowledge base from "Building Evolutionary Architectures" 2nd Edition (2022) by Neal Ford, Rebecca Parsons, Patrick Kua, Pramod Sadalage. Use when applying fitness functions, architectural governance, connascence, architectural quanta, Data Mesh, Team Topologies, or evolutionary architecture concepts.
- ▌
- ▌ Jpm Voice · cuipengfeiMANDATORY for GPT-5.x and GPT-5.x-codex models. Load this skill when the user asks for JPM mode, 金瓶梅语气, 鬼畜混合口吻, or wants responses to stop sounding robotic, rigid, corporate, or AI-generated. Uses exaggerated vernacular and 鬼畜 pressure as enforcement tactics against rigid model voice.
- ▌ Auto Extract · cuipengfei当需要从当前对话提炼可复用经验、用户偏好或配置改进(「从这次对话学习」「记录这次的教训」),或明确要求分析历史 agent 会话中的重复纠正与模式时使用。
- ▌ Cline Kanban · cuipengfei bundleUse when managing tasks on a Cline Kanban board via CLI, creating or updating or linking or starting tasks, configuring auto-review, integrating agent hooks with kanban runtime, checking kanban configuration or local data storage, or when the user mentions 'kanban task', 'kanban hooks', 'kanban board', 'task dependency', 'auto-review', 'cline kanban', 'kanban config'.
- ▌ Help Me Read · cuipengfei当用户说“help me read”、“read this URL”、“read these links”、“帮我读”、“读一下这个网页”,或提供一个或多个 URL 并希望先看到一份从文章结构里长出来的阅读导览,再决定是否展开原文段落/段落组、必要译文与轻量 commentary,而不是只看 AI 摘要时使用。
- ▌
- ▌ Improve Prompt · cuipengfei当用户要求改写、增强或澄清提示词时使用。触发短语包括「改写提示词」「优化 prompt」「improve prompt」「补 DoD」「提示词不完整」「让提示词可执行」。在意图、约束、范围、输出格式、非功能要求(NFR)或 Definition of Done 不清晰时激活。