open-skills

当用户说「修复这个 bug [URL]」「帮我修复 [URL]」「jira-fix [URL]」「自动修复 [URL]」「强制修复 [URL]」「继续修复」「从上次继续」时触发。适用于从 Jira 链接出发、对单个 bug 进行端到端修复的场景。

合并纪律：合并动作（glab/gh mr/pr merge）前必须加载——rebase/冲突预检（Part A，最先）+ 覆盖率门控（Part B，test-coverage-analyzer）+ tip 钉死（Part C，--sha 钉死 + 祖先校验），防盲目合入已移动的目标分支 + push→merge 竞态致 archive/修复未入目标分支。触发词：「合并 tip」「merge tip」「合并纪律」「push 后合并」「archive 合入」「合并前门控」「rebase 检查」「冲突预检」「合并前 rebase」 / merge discipline, rebase pre-check, coverage gate, post-push merge check. 被 opsx-jira-fix-workflow / opsx-solve-workflow / jira-fix-workflow frontmatter dependencies 强依赖。

当用户说"opsx-jira-fix"、"OpenSpec Jira 修复"、"规范化修复 Jira"、"opsx修复Jira"、"Jira OpenSpec 修复"、"opsx自动修复Jira"、"用OpenSpec修复Jira"或"opsx-jira-fix-workflow"时触发。适用于从 Jira issue 出发，并需要将根因、行为变更、修复计划、验证和归档沉淀到 OpenSpec artifacts 的端到端 Bug 修复。

当用户说"opsx解决"、"OpenSpec解决"、"规范化解决"、"创建OpenSpec变更"、"创建opsx变更"、"用OpenSpec分析"、"用OpenSpec修复"、"opsx自动解决"、"OpenSpec自动解决"、"opsx-solve"或"opsx-solve-workflow"时触发。适用于需要将分析、方案、计划、实现、验证和归档沉淀到OpenSpec artifacts的功能开发、Bug修复、重构和复杂工程任务。

Use when releasing a Git repository version — tagging, merging release branches into main, resolving conflicts, or syncing changes between release branches. Handles ambiguous tag naming (v-prefix vs plain), unknown main branch (master/main/develop), ff-only merge method (forces cherry-pick strategy), protected-branch rules, cross-release hash-sensitive sync, and MR/PR extra-file cleanup. Single or multi-repo. Triggers — 「发版」「打tag」「发布版本」「版本发布」「release流程」「git-release」「multi-repo release」「release同步到另一个release」 / release a version, tag and merge, finish release branch.

Browser runtime debugging toolkit — guides AI to prioritize browser DevTools and CDP-based tools for runtime inspection and control when debugging UI/CSS/DOM layout, frontend interaction, and rendering issues. Two CDP channels: chrome-devtools-mcp (inspection, DevTools panels) and the web-access skill's CDP Proxy (control + verify, carries login state, curl-scriptable; runtime-checked, aborts with an install hint if web-access is absent). Also covers Playwright. Triggers: 「浏览器调试」「UI 调试」「DOM 检查」「CSS 调试」「页面布局问题」「前端运行时调试」「chrome devtools」「CDP 调试」「登录态调试」 / browser debug, devtools, dom inspect, css debug, runtime debugging, login-state debug.

Effective web research discipline for AI agents — route first, then research with rigor. Step 0 triages whether a question needs external web lookup vs is answerable from the internal codebase/docs; when external, default mode auto-applies 4 maxims (official sources first, check recency, cross-validate non-trivial claims, skip content farms) so every web search is effective and credible; on explicit 'strict/deep research' requests it runs a full 7-dimension source-credibility evaluation (CRAAP + E-E-A-T merged) and emits an auditable report. When the static layer (WebSearch/WebFetch/curl) can't reach the content (login-required / JS-rendered / anti-scraping platforms like 小红书·公众号), it escalates to the real browser via the web-access skill (CDP, carries login state). Triggers: 「web 调研」「外部调研」「查资料」「有效调研」「严格调研」「深度调研」「严格查证」「登录态访问」「动态页面」「反爬」「这个库/框架怎么用」「有没有漏洞」「best practice 是什么」 / web research, look up, investigate, strict research, deep research, fact-check, login-state access, dynamic page, anti-scraping. Do NOT u

当用户说「批量修复」「批量 jira-fix」「jira-fix-batch」「批量修复多个 Jira」「批量修复以下 bug」时触发。适用于需要对多个 Jira issue 进行批量端到端修复的编排场景。

当用户说「opsx 批量修复」「批量 opsx-jira-fix」「opsx-jira-fix-batch」「批量 OpenSpec Jira 修复」时触发。适用于需要对多个 Jira issue 进行批量端到端修复并将关系判断沉淀到 OpenSpec artifacts 的编排场景。

Ensure the current project has a proper test suite — detect tech stack & framework, scaffold if needed, generate unit tests (required, logic code only), and optionally generate/run E2E tests. Triggers when user says "ensure-tests", 「补全测试」「生成测试」「确保测试」「补充单元测试」「添加单元测试」「检查测试覆盖」 (complete tests / generate tests / ensure tests / add unit tests / check test coverage). Also callable by solve-workflow and opsx-solve-workflow in their execution-stage test steps.

Node version discipline — before running tsc / eslint / build / test / install in any Node project, align the Node version to the project's declared version (.nvmrc / .node-version / .tool-versions / volta / engines.node) and disclose it in the verification report, preventing false-pass / false-fail when the host default Node mismatches the project-pinned version; when the project declares no Node version at all, ask the user for the version and offer to persist a declaration file (.nvmrc etc.) so future sessions/CI/collaborators auto-align. Use this skill whenever the user is about to run a version-sensitive Node command, or mentions Node version mismatch / wrong Node / version conflict / 切换 node 版本. 中文触发词「node 版本对齐」「版本对齐」「nvm 对齐」「.nvmrc」「Node 版本纪律」「对齐 Node 版本」「切换 node 版本」「切到指定 node 版本」「node 版本不对」「node 版本不一致」「版本不匹配」「初始化 node 版本管理」「为工程增加 node 版本声明」「给工程加 .nvmrc」「没有 .nvmrc 怎么办」「无 .nvmrc 时补一个」「工程没有 node 版本管理」, English aliases node version discipline, nvm use, align node version, switch node version, add nvmrc, i

Methodology for escalating from static code analysis to runtime evidence gathering when static analysis hits a wall. Make sure to use this skill whenever you are debugging and static code analysis alone has not confirmed the root cause with high confidence — it covers the full evidence-gathering lifecycle: escalation decision (scientific debugging, Zeller), instrumentation design (observability triad and structured logging — NOT shotgun console.log), reproduction guidance (MRE/MCVE and deterministic reproduction), evidence analysis (evidence hierarchy: runtime observation stronger than log stronger than code inference), confidence gating (Bayesian calibration — when is evidence strong enough to act), escape hatch (when runtime cannot penetrate platform/host layers — git bisect, web research), and fix verification (before/after comparison, not just reading the fixed code). Use this skill when root-cause confidence is low or fuzzy after static analysis, when you need to add logging or instrumentation to observe

当用户说"明确问题"、"分析问题"、"探索方案"、"审查方案"、"制定计划"、"执行计划"、"检查验证"、"复盘改进"、"回顾总结"，或"继续分析"、"深入分析"、"修改方案"、"完善方案"、"优化方案"、"更新计划"、"修订计划"、"修改计划"，或"自动模式"、"自动分析"、"自动解决"时触发。适用于 bug 修复、代码重构、功能开发等需系统性分析的复杂任务。

Cross-platform environment capability discovery for workflow skills: scan once at startup for available enhancement skills/agents (debugging, web research, planning, TDD, build-fix, verification, branch management, etc.), match by keywords, and invoke them as progressive enhancements without changing stage gates or tool constraints. Weak reference by default — workflows normally reference it without declaring it in dependencies; workflows that do declare it (e.g. solve-workflow) get availability guaranteed by their prerequisite check. Load it when a workflow references it; when unavailable and not declared, skip silently.

Hard discipline for asking the user clarifying questions: ask exactly ONE most critical question per round (priority: purpose → constraints → success criteria), and resolve multiple unknowns across MULTIPLE rounds — each round's question refined by the previous answer (not 'only one question for the whole session'). Also carries the investigation-first principle. Referenced via frontmatter dependencies by workflow skills (solve-workflow, opsx-solve-workflow, jira-fix-workflow, opsx-jira-fix-workflow, perf-workflow); load when a workflow delegates user-questioning guidance.

External-research routing for confirmed code problems: triage whether the root cause is internal / external / hybrid, run a known-issue quick search before deep root-causing (platform silent failures, nested host runtimes, no code-level suspects), and evaluate industry-wide hard limits. Delegates all WebSearch discipline to effective-web-research. Referenced via frontmatter dependencies by workflow skills (solve-workflow, opsx-solve-workflow, jira-fix-workflow, opsx-jira-fix-workflow); load during the technical-analysis stage.

性能问题分析与优化工作流，共六阶段。触发词均以「性能」开头：性能分析、性能证据、性能定位、性能假设、性能监控、性能优化、性能验证、性能深入。当用户说上述词或使用「触发词： 具体描述」形式时，进入本工作流或对应阶段。

Shared lifecycle contract for manual/auto modes in PDCA-style workflow skills: auto mode always reverts to manual on completion or interruption, re-entering auto requires an explicit trigger, implicit continuation never re-activates it, and batch orchestrators pass mode explicitly. Referenced via frontmatter dependencies by solve-workflow, opsx-solve-workflow, jira-fix-workflow, opsx-jira-fix-workflow; load when a workflow delegates its mode rules.

Strategic thinking framework — analyze whether something is worth doing, what the risks are, and what the long-term impact will be. Use when user says 「战略思考」「战略分析」「从战略角度看」「战略视角」「帮我想清楚这件事」「这件事值不值得做」「从更高视角分析」 / think-big, strategic thinking, strategy. Not for execution-level planning (that's tactical).

Structured retrospective and knowledge sediment for completed work (the Act phase of PDCA). Run a disciplined post-work review (Keep-Problem-Try / Stop-Start-Continue / After-Action Review by context), judge which lessons are worth solidifying, pick the right carrier, validate the sediment gets reused, and feed improvements back. Use whenever finished work needs reflection — as the review stage of solve-workflow / opsx-solve-workflow, or standalone on any completed task. Triggers — 「回顾总结」「复盘」「经验沉淀」「总结经验教训」「改进闭环」「沉淀规则」「项目复盘」 / learn and improve, lessons learned, retrospective, post-mortem, debrief.

Implement tasks from an OpenSpec change. Use when the user wants to start implementing, continue implementation, or work through tasks.

Archive a completed change in the experimental workflow. Use when the user wants to finalize and archive a change after implementation is complete.

Archive multiple completed changes at once. Use when archiving several parallel changes.

Continue working on an OpenSpec change by creating the next artifact. Use when the user wants to progress their change, create the next artifact, or continue their workflow.

Enter 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.

Fast-forward through OpenSpec artifact creation. Use when the user wants to quickly create all artifacts needed for implementation without stepping through each one individually.

Start a new OpenSpec change using the experimental artifact workflow. Use when the user wants to create a new feature, fix, or modification with a structured step-by-step approach.

Guided onboarding for OpenSpec - walk through a complete workflow cycle with narration and real codebase work.

Sync delta specs from a change to main specs. Use when the user wants to update main specs with changes from a delta spec, without archiving the change.

Verify implementation matches change artifacts. Use when the user wants to validate that implementation is complete, correct, and coherent before archiving.

当 Git merge 或 rebase 过程中出现代码冲突时使用，尤其适用于 AI 自动解冲突容易出错（取错侧、丢失重构、还原旧版代码）、需要语义分析和逻辑验证的场景。也适用于 rebase 多轮停止需跨轮聚合冲突文件的情况。触发词：解冲突、处理冲突、git-conflict-resolve、解决 merge 冲突、解决 rebase 冲突、conflict resolve、冲突解决。

TypeScript type-checking workflow. Use this skill whenever the user works with TypeScript and mentions type errors, type checking, or says "类型检查" (type check), "type-check", "tsc", "检查类型" — even if they don't explicitly ask for a "type check" but are encountering compilation or type issues in a TypeScript project.

Use when 用户想写公众号、知乎技术文章、自媒体内容，或基于已有技术文档改写为可发布文章时。

When a bug involves a named third-party library/framework, or symptoms correlate strongly with a platform/library-specific behavior, evaluate 'upgrade the dependency' BEFORE piling on local workarounds. Walks through a 4-step decision (is it upstream? → check Changelog/Release Notes → upgrade at low semver risk → only workaround if upgrade is infeasible), the upgrade engineering discipline (package-manager consistency, post-upgrade verification chain, semver risk, dedup check), and a result table. Use this skill whenever a bug smells like an upstream library issue — symptoms tied to a specific framework version, silent failures where code looks correct, workarounds that keep growing, or platform-specific rendering bugs that CSS cannot fix. Triggers — 「升级依赖」「依赖升级修复」「这个 bug 升级依赖能解决吗」「查 changelog 修复」「优先升级依赖」「升级依赖还是堆 workaround」 / dependency upgrade, upgrade to fix bug, changelog fix, avoid workaround, upgrade library to fix.

Authoritative framework for reviewing the design quality of proposed code changes — before implementation. Evaluates code-level metrics (accidental complexity, coupling via Myers/Connascence, cohesion, change amplification, tech debt, cyclomatic/cognitive complexity, Law of Demeter) and architecture-level quality attributes (testability, modularity, reliability, scalability, dependency direction via ISO 25010 + Clean Architecture + SDP) plus a security pass (OWASP Top 10). Use this skill whenever a proposed solution involves code changes and you need to assess whether the code is well-designed — not just whether it works, but whether it is maintainable, testable, and does not introduce design debt. Triggers — 「代码审查」「代码设计审查」「代码设计质量」「审查代码设计」「代码架构审查」「设计质量评估」「代码质量评审」「这个代码设计合理吗」「耦合度审查」「代码可维护性」 / code design review, code architecture review, design quality assessment, coupling analysis, maintainability review.

Hybrid app (native shell + WebView/WKWebView/Electron + H5) full-stack debugging methodology — forces analysis across four layers (web runtime, native-web bridge, native config, platform runtime) instead of stopping at a single layer, preventing whack-a-mole surface fixes. Use when debugging hybrid app UI/theme/behavior issues, especially platform discrepancies (one platform reproduces, the other does not), repeated single-end fix failures, or silent failures involving native-H5 communication, theme switching, or platform runtime behavior. Triggers — 「hybrid 调试」「跨端调试」「全链路调试」「平台差异调试」「WebView 问题」「WKWebView 问题」「native 和 H5 交互问题」「单端修复失败」「跨端主题问题」「跨端样式不一致」 / hybrid debug, cross-platform debug, webview issue, native-web bridge, full-stack debugging, RN/Electron debugging.

Authoritative framework for reviewing any proposed solution (code, config, process, tooling, or architecture decision) before implementation. Covers four core dimensions (effectiveness, side-effects/risks, feasibility, spec compliance) plus five strategic dimensions that catch decision-level failures the core four miss: reversibility calibration (Bezos one-way/two-way doors), failure-mode analysis (FMEA), operability (SRE Production Readiness Review), cost-vs-value (CBAM/WSJF), and team cognitive fit (Team Topologies). Use this skill whenever you need to review, validate, stress-test, or approve a proposed solution, plan, design doc, or approach — before any code is written or change is applied. Triggers — 「审查方案」「方案审查」「评估方案」「方案评估」「review proposal」「方案可行吗」「这个方案靠谱吗」「方案有没有风险」「方案评审」「设计评审」「决策评审」 / solution review, proposal review, design review, decision review, stress-test a plan, validate an approach.

Use when a complex problem seems unsolvable and needs digging to the essence — diagnosis only, not fixing. Symptoms include massive logs with no root cause surfaced, complex crashes, architectural essence analysis, requirement authenticity verification, systemic chronic issues, information overload, disconnect between symptoms and essence, multiple intertwined factors, or premature-convergence tendencies — even when the user never explicitly says "diagnose". Also triggers on 本质诊断, 根因诊断, 深度诊断, 诊断问题本质, 梳理问题逻辑, 证据链分析, 逻辑链分析, essence-diagnosis.

Universal file/container unpacker for AI agents. Use whenever the user has a specific named file that is secretly a container of other files (an .exe installer, .docx document, .apk app, .asar Electron bundle, .dmg disk image, .deb/.rpm package, .jar/.war, .iso, .pkg, .whl, or any archive) and wants to extract or inspect its inner contents — even when they don't say 'unbox' explicitly. Cracks ~30 'box-like' formats into clean directories with a structured manifest: standard archives (zip/tar/7z/rar/zst/lz4), ZIP containers (docx/xlsx/pptx/jar/war/apk/ipa/whl), system packages (deb/rpm/msi/cab), installers & disk images (NSIS exe/Inno Setup/dmg/iso/pkg), and chained nested containers (exe→7z→asar, pkg→xar→cpio). Cross-platform (mac/linux/windows) via 7z as primary tool with graceful fallbacks. Triggers: 「拆箱」「开箱」「解压文件」「拆包」「解压任意文件」「提取容器内容」「拆开[文件]」「看[文件]里面」 / unbox, unpack any file, extract container, what's inside this file. Do NOT use for: writing extraction code/scripts, creating or editing Office documents, i

Read Jira issue data from local cache or API. Triggers when user says "jira-read [JIRA-ID]", 「读取 Jira」「查看 Jira」「下载 Jira」 (read/view/download Jira), or needs to fetch Jira issue data. Requires $JIRA_CACHE_DIR (e.g. ~/.cache/jira).