Use when the user asks what they are not seeing, what they forgot, what could go wrong, or what an expert would say about a plan, decision, document, code change, or setup they already have. Triggers: "what am I missing", "what don't I know", "poke holes in this", "what would an expert ask", "что я упускаю", "что я не вижу", or /clueless-blindspots. One-shot review; does not change the always-on level.
中文简介正在准备中。当前显示原文说明。
查看安装方法安装前可以查看交给 AI 代理的指令内容。
Review the thing in front of you as the expert the user doesn't have. Find what they didn't know to ask. Output is a list, nothing else.
One line per finding, highest stakes first:
<tag> <what they missed>. <what to do about it>.
Tags:
irreversible: data loss, money gone, legal exposure, locked-out users, missed deadline. Always first.assumed: a decision baked in without being named. Say what the alternative was.unasked: a question a professional would have asked. Answer it with the safe default.unverified: a claim, number, or step nobody has checked. Say how to check it.placeholder: a value that will ship as-is if nobody replaces it.❌ "You might want to consider whether the backup strategy fully covers your needs."
✅ irreversible: iCloud Photos is sync, not backup; delete on the phone deletes everywhere. Keep the Time Machine drive as the real copy.
✅ assumed: SQLite as the database. Fine under ~100 users; if you expect more, say so now, migrating later is a weekend.
✅ unasked: which state you're in. State income tax is a separate filing; assumed a state that has one.
✅ unverified: the 25–30% tax set-aside. Ask the CPA to confirm against last year's return.
✅ placeholder: session secret is still "change-me". Set SESSION_SECRET in the environment before the site goes live, or the server should refuse to start.
End with one line: <N> irreversible, <M> assumed, <K> unasked.
If there is nothing: Nothing you're not seeing. Go.
Lists only, does not fix. Stakes and blind spots only: style, taste, and "could be cleaner" are out of scope. Never invents risk to look thorough; an empty category is an empty category.
还没有评价。欢迎分享你的使用感受。
无需登录即可发表。欢迎分享使用感受、提问，或说明在哪些环境下无法运行。
还没有评论。来写第一条吧。
Use AtomLane to compile and execute safe atomic parallel plans on macOS and native Windows Preview for worthwhile independent argv tasks, dependency DAGs, supported platform entrypoints, or Apple-silicon operators. Use at task start or an execution boundary when structured local work may contain two or more worthwhile units; skip plain answers, one quick command, and work whose effects cannot be safely bounded.
中文简介正在准备中。当前显示原文说明。
Register a deferred decision in the debt registry. Trigger by judgment, not a marker scan, whenever a future reader would ask "why this way?": an unmade decision, stub, loosened type, bypassed check, swallowed error, a default picked "for now", or a TODO/FIXME/HACK/XXX marker. Trigger immediately whenever you defer work, or when the user invokes $add. Over-register freely; the developer drops with "drop A", "drop A,C", or "drop all".
中文简介正在准备中。当前显示原文说明。
ADK 框架适配层。为 LangChain / EINO / AutoGen / AgentScope / CrewAI 提供框架特定的 代码模板、惯用模式、API 映射和项目结构，供 agent-dev-workshop Phase 5 代码生成使用。 每个框架 reference 文件标注 verified_date 用于版本锁定。
中文简介正在准备中。当前显示原文说明。
中文调试修复技能。用于报错、测试失败、页面异常、功能不符合预期、需要定位根因并做最小修复时。触发语包括"进入调试模式""帮我修问题""报错了""测试失败""页面坏了""找根因"。
中文简介正在准备中。当前显示原文说明。
交互式 AI Agent 开发工作坊：通过 6 阶段深度协作对话，引导用户完成 Agent 需求分析、架构设计、 工具定义、Prompt 与编排设计、代码生成、验证迭代，产出可直接运行的 Agent 项目。 框架无关设计优先，支持 LangChain / EINO / AutoGen / AgentScope / CrewAI 等 ADK 框架。
中文简介正在准备中。当前显示原文说明。
中文漂移审计技能。用于项目或学习过程变乱、上下文漂移、任务分叉、多个方案冲突、命名不一致、Codex 可能顺手改多了时。触发语包括"漂移检查""感觉跑偏了""项目变乱了""检查是否失控""分叉太多""上下文漂移"。
中文简介正在准备中。当前显示原文说明。