Coding standards, frameworks, refactoring, and code review (16K skills)
TypeScript (tsc, 6.0 / 7.0 Go 製ネイティブ版) リファレンス。 tsconfig.json compilerOptions, strict, moduleResolution, project references (tsc -b), generics, conditional / mapped / template literal types, utility types, narrowing, .d.ts, checkJs, ESM / CJS interop, --checkers / --builders。
The English summary is being prepared. Showing the original description.
React Native (Meta 製クロスプラットフォームネイティブアプリフレームワーク, react-native 0.87 / 0.88) の リファレンス。View, Text, FlatList, Pressable, TextInput, Animated, StyleSheet, Flexbox, Platform, New Architecture (Fabric, TurboModules, Codegen), Hermes, Metro, Fast Refresh, TypeScript, React Native DevTools, Android / iOS 統合、upgrade (Upgrade Helper)・0.84〜0.88 差分。
The English summary is being prepared. Showing the original description.
Google Gemini API (Gemini Developer API, ai.google.dev) の Interactions API (v1 / v1beta), generateContent, Google GenAI SDK (google-genai / @google/genai), Live API, ephemeral tokens, Managed Agents, Deep Research, function calling, Google Search grounding, File Search, Files API, thought signatures, Nano Banana, Veo, Lyria, Batch API, Flex / Priority inference, OpenAI 互換。
The English summary is being prepared. Showing the original description.
Model Context Protocol (MCP 仕様 2026-07-28, modelcontextprotocol.io) リファレンス。 JSON-RPC, server/discover, tools/call, subscriptions/listen, MRTR (InputRequiredResult), CacheableResult, Streamable HTTP, CIMD, MCP Apps (ui://), MCP Registry (server.json, MCPB), tasks 拡張。ホスト設定・SDK は別スキル。
The English summary is being prepared. Showing the original description.
Kotlin Multiplatform (KMP) / Compose Multiplatform (JetBrains, Kotlin) の公式リファレンス。 expect/actual, source set hierarchy, Multiplatform Gradle DSL, org.jetbrains.compose, Compose Resources (Res), Compose for iOS / Compose Desktop (Window, Tray, MenuBar, nativeDistributions) / Web, iOS 統合 (SwiftPM export, CocoaPods), Navigation / ViewModel / Lifecycle, Compose Hot Reload, Kotlin Toolchain, CI (GitHub Actions / TeamCity), Maven Central / npm 公開, AGP 9 移行, 互換性表。
The English summary is being prepared. Showing the original description.
当手上有一个零散的读书会、学生分享、讲座或社群活动想法，想知道下一步该怎么长时使用。先按留下的痕迹（办过几次、来了几个人、留下了什么）判出它实际在哪儿，而不是按它叫什么；再用带证据清单的闸门决定是留在原地、修改后再试、往深处走（Workshop）还是往广处走（公共活动）。深度与广度是两个独立的轴，不是一条梯子。产出各阶段的标准表单与交接，不替使用者编造学校制度、人员、预算或未给出的活动形式定义。
The English summary is being prepared. Showing the original description.
当使用者要做苹果发布会风格的 PPT、把演讲稿做成幻灯片、把现有 PPT 改成发布会风格，或者要做讲座、课堂、学术报告、答辩的投影片时使用。输入可以是一句描述、一份演讲稿或讲稿、一份现有 .pptx。先判断场合（发布会／答辩／课堂／讲堂），再分两步：先产出 deck-outline.md（拍点、逐片清单、出场顺序、出处、舍弃项）等确认，再产出单文件 deck.html（1920×1080、讲到才出现的逐步揭示、跨片连续过渡、讲者备注、原生公式、可打印 PDF），需要时导出可编辑 .pptx（动画、调暗、点名与 Morph 平滑切换都带过去），课堂与讲堂另出讲义。核心规矩：一片一个概念、不用项目符号、字号不低于 40px、片子不重复讲稿、数字与图片都要有出处。片上默认用名词短语、不用问句，发布会片上没有问号；课堂与讲堂的证据片改用一句主张，问号只给两处：真要听众作答、答案随后揭晓的提问片，以及下一张片立刻用一句结论回答的引导问句（一节课一两处）。片上文字另按内嵌的中文写作规则走。不使用任何厂商的商标、产品渲染图或活动视觉。
The English summary is being prepared. Showing the original description.
当使用者说“按我的模型写代码”“把方程实现成 Python/MATLAB/R”“调试这段建模代码”“补验证和灵敏度分析”“让结果可复现”时使用。只实现学生已确认的模型与数据契约，先写实现合同、手算小例、不变量和验收矩阵，再按读取、变换、求解、验证、敏感性与导出拆分；执行后记录输入快照、种子、环境、求解设置和输出，不静默修数据、不硬编码结果、不替学生发明模型或伪造数据。
The English summary is being prepared. Showing the original description.
当使用者给出「说话人名字 + 说话内容 + 主色」，说"做一张车队无线电那种语录卡"、"生成 XX RADIO 的图"、"把这句话做成无线电字幕卡"、"照那个红色波形的格式来一张"时使用。产出单文件 HTML，800×1000 竖版：顶部是名字（主色）加 RADIO（白色）的粗斜体标题、中间一条主色音频波形加通栏横线、下方是主色粗斜体全大写的引号语录、底边一条主色横条，底色固定为近黑。使用者没给图标就不放图标。不生成任何车队或厂商的徽标，不伪造真人说过的话。
The English summary is being prepared. Showing the original description.
当使用者说"把这些照片排成一版"、"做一张照片展示海报"、"仿报纸摄影版那种排版"、"这组照片怎么摆"、"做个 photo essay 的版面"时使用。输入一组照片加署名，分两步：先产出 spread-plan.md（分带、定主次、排序、每带的图注与署名、被舍弃的照片及原因），等确认后再产出单文件 spread.html（照片按需缩放并内联，可打印成 PDF）。这个 skill 不提供固定版式——它给的是带状构图的语法与协议，具体谁大谁小、分几带、怎么排由模型和使用者判断。每张照片必须有署名，分辨率不足会被拦下。只想把自己拍的 3–9 张按原顺序拼成一张长图、网格或一页 PDF，不要分带主次和署名时，用 photo-series-layout。
The English summary is being prepared. Showing the original description.
已废弃，改用 bento-infographic（发布会 bento 信息图）。原用途：当使用者贴出产品资料、发布稿、规格表、卖点清单或网页摘录，说"做一张产品总结面板"、"生成发布会那种 summary slide"、"把这些卖点做成一张图"、"bento 风格的产品总览"时使用。分两步：先产出 panel-brief.md（定位、核心卖点、支撑卖点、数字锚点、信息分组、每条的来源或未确认标记、版式建议、舍弃项与原因），等使用者确认后再产出单文件 panel.html（16:9 横版、内联 CSS、bento 卡片布局、light/dark 双主题、适合截图）。只借鉴信息层级与面板结构，不复制任何厂商的资产、商标、截图或图标。不编造参数，无出处的数字标注或省略。
The English summary is being prepared. Showing the original description.
当使用者说"做一张苹果发布会那种总结图""bento 图""把卖点、功能或研究成果排成一张格子图""产品亮点一图看完""做成竖版发小红书"时使用。先出内容清单和线框预览，确认后由脚本排版，用本机 Chrome 渲染成 PNG：主图居中，四周十几个圆角格，字号、格缝、圆角按苹果样图的实测比例换算，横版、竖版、任意比例都能排，渲染后自动跑 19 项自检。研究成果走研究模式：测量值带不确定度、样本数与条件，可放一张小图表。图上的数字必须能指回材料，没有出处的不放，或在图上标明未确认。不用苹果的商标、产品图、SF Symbols 和字体文件；不生成图片，图片由使用者提供。
The English summary is being prepared. Showing the original description.
当拿到一段 AI 回答、准备直接复制去用之前使用。把回答拆成可检验的声明（链接、库名与版本、论文 DOI、命令行参数、API 名、数字、引文），按"用错了会炸多大"排序，为每条配一条最直接的核验路径；产出声明台账、核验计划和"现在可以用 / 必须先核实"两栏。专门防住引用了不存在的库、论文或命令这种直接复制就会出事的情况；包名存在不等于可信（幻觉出的名字可能已被人抢注），DOI 能解析也要标题对得上。不替使用者背书，核不了就写核不了。与 ai-answer-triage 的区别：那边只分级、定先查哪几条，这里逐条给核验路径；要判断一整段 AI 代码能不能跑，用 ai-code-onboarding-checklist。
The English summary is being prepared. Showing the original description.
当拿到一段 AI 生成的代码，还没决定要不要信、要不要跑、要不要往上继续加东西时使用。先做一遍客观体检：行数与嵌套深度、有没有测试、有没有硬编码密钥或遗留 TODO、依赖是不是真实存在、有没有危险调用（eval/exec、shell=True、递归删除、反序列化、提权、外发网络，按正文给的 grep 模式逐类查）；再给风险评级和下一步动作。老实写明它查不了算法正确性——体检通过不等于代码是对的。回答里文字、事实和代码混在一起时先用 ai-answer-triage 分级；只审相对已有代码的一份改动用 ai-diff-review-protocol。
The English summary is being prepared. Showing the original description.
当想知道一段 AI 回答"哪些能直接用、哪些必须先验证、哪些只能当建议"时使用。按声明类型分四级——低风险草稿材料、观点与策略建议、可查证的事实声明、会改变状态的可执行指令——每级配一句默认处置动作；再按影响半径排序，给出最小的下一步检查。可执行类一律标"不能直接执行；先做代码审阅/沙箱测试/回滚准备"，其中破坏性或提权的命令排在最前。教的是一套可复用的复查习惯，不是替使用者把所有内容都核实一遍：本 skill 只分级、排序、定处置，要逐条核实走 ai-output-fact-checker，整段代码要做入库体检走 ai-code-onboarding-checklist。
The English summary is being prepared. Showing the original description.
The runbook a triage agent follows on each run over the super-admin attention dock, through the attention MCP tool. Use when a scheduled attention-triage routine wakes, when asked to triage, clean up or shrink the attention dock or admin alerts, or when the dock's card is over budget.
The English summary is being prepared. Showing the original description.
Check the latest Claude (Anthropic), OpenAI Codex, and Google Gemini model releases from official primary sources, update model selections in this dotfiles repo, and update the Claude Code, Codex, and Antigravity (agy) CLIs through their configured managers. Every run checks both the model settings and CLI for the selected provider, then scans the invoking repository for hardcoded model IDs. Use when the user asks to "モデル更新", "モデルを最新に", "最新モデル確認", "Codex/Claude Code/agy本体の更新", "model bump", "update models", or "update agent CLIs". Do NOT use for one-off model selection in a single conversation, general model questions, or unrelated package updates.
The English summary is being prepared. Showing the original description.
Delegate writing predictable code to the `code-write` command, which has a cheaper writer model write the target file directly so the generated code never enters your context. Use for tests, scaffolding, fixtures, boilerplate, or a new file that copies an existing pattern, and when the user says "code write", "生成を委譲", or "雛形を書かせて". Do NOT use for design decisions, changes to public contracts or shared modules, edits inside an existing file that needs judgment, or logic with subtle correctness requirements.
The English summary is being prepared. Showing the original description.
Operate Orca-managed worktrees, folder contexts, terminals, repos, automations, artifacts, skill sharing, worktree comments, and Orca's embedded browser through the `orca` CLI. Use when the user says "$orca-cli", "Orca worktree", "child worktree", "spawn codex/claude in a worktree", "read/wait/send Orca terminal", "Orca browser", "orca artifacts", or "share skills". Do NOT use for handing a task to another session; use my-handoff. Prefer it over raw git worktree, ad hoc PTYs, or Computer Use when Orca state is involved. Use Computer Use only for external windows or desktop UI that needs OS-level control, and Playwright or CDP for external pages.
The English summary is being prepared. Showing the original description.
思考フレームワークに基づく対話・レビュー・執筆スキル。「私の視点で」「私の考え方で」をプロンプトに含めると適用される。
The English summary is being prepared. Showing the original description.
Discover and inventory shadow API endpoints that operate outside documented specifications using traffic analysis, code scanning, and API discovery platforms.
The English summary is being prepared. Showing the original description.
Deobfuscates malicious JavaScript code used in web-based attacks, phishing pages, and dropper scripts by reversing encoding layers, eval chains, string manipulation, and control flow obfuscation to reveal the original malicious logic. Activates for requests involving JavaScript malware analysis, script deobfuscation, web skimmer analysis, or obfuscated dropper investigation.
The English summary is being prepared. Showing the original description.
Detect and prevent API enumeration attacks including BOLA and IDOR exploitation by monitoring sequential identifier access patterns and authorization failures.
The English summary is being prepared. Showing the original description.
Configure secure OAuth 2.0 authorization flows including Authorization Code with PKCE, Client Credentials, and Device Authorization Grant. This skill covers flow selection, PKCE implementation, token
The English summary is being prepared. Showing the original description.