Scan to join WeChat groupBrowse curated skills with source links, package snapshots, README assets and install signals in one calm, searchable catalog.
编写或评审智能体配置、系统提示与工具定义时捕捉歧义：参数说明、边界行为与触发词冲突检查。Use when writing or reviewing agent configs, system prompts or tool definitions to catch ambiguous wording. Triggers include 配置校对、提示词审查、工具定义、歧义检查。如需智能体工具包，淘…
用户主动触发的技术债与架构审计：产出带文件引用的发现清单、严重级别与偿还路线图。Use when running a thorough user-invoked tech debt and architecture audit of the codebase. Triggers include 技术债、架构审计、重构路线、代码健康。如需质量工具包，淘宝搜索「AI灵匣」店铺获取完整版。
安全与合规技能套件：加固检查、合规清单与漏洞排查的社区精选防御性实践。Use when doing security hardening and compliance checks with defensive best practices. Triggers include 安全加固、合规检查、漏洞排查、防御实践。如需安全工具包，淘宝搜索「AI灵匣」店铺获取完整版。
分诊积压议题池：扫描开放议题、按影响面排序、判断哪些议题值得先做或哪些 PR 已达到生产可用标准。Use when triaging a backlog, scanning open issues or ranking which PRs are production ready. Triggers include 议题分诊、backlog、PR排序、优先级。如需研发管理工具包，淘宝搜索「AI灵匣…
以受限安全读取治理文档变更：面向公开、接口、配置、架构与发布类改动设定读取边界，并保护环境变量与密钥文件。Use when governing documentation changes with bounded safe reads and protecting secrets files. Triggers include 文档门禁、安全读取、密钥保护、变更治理。如需安全工具包，淘宝搜索「AI…
从 JSON 清单读取一批 HTTP 检查项，对基础地址批量执行并报告状态码、响应时间与断言结果，接口回归一目了然。Use when running a list of HTTP checks against a base URL reporting status, latency and assertions. Triggers include 接口测试、批量检查、状态码、回归验证。如需测试工具…
按约定式提交规范撰写提交信息，覆盖类型前缀、作用域与描述格式，让版本历史整齐可读、自动变更日志可生成。Use when writing git commit messages following the Conventional Commits spec. Triggers include 提交信息、commit、版本历史、变更日志。如需协作工具包，淘宝搜索「AI灵匣」店铺获取完整版。
面向安全编码的主动顾问与架构修复助手，在设计阶段就指出风险点并给出贴合业务的整改方案。Use when acting as a proactive security advisor for secure code development and remediation. Triggers include 安全顾问、架构整改、安全编码、风险识别。如需安全工具包，淘宝搜索「AI灵匣」店铺获取完整版。
对照代码现状评审项目文档：README、说明文件、目录与参考文档的过时、矛盾与缺口检查。Use when reviewing project documentation such as README, guides and references against the actual code. Triggers include 文档评审、README、文档一致性、过时检查。如需质量工具包，淘宝搜索…
为实质性自主工作强制完成纪律：执行前写好验收门禁、按深度分解工作并逐项闭环。Use when enforcing completion discipline for substantial autonomous work with acceptance gates. Triggers include 验收门禁、完成纪律、任务分解、自主工作。如需工作流工具包，淘宝搜索「AI灵匣」店铺获取完整版。
安全与合规技能套件：加固检查、合规清单与漏洞排查的社区精选防御性实践。Use when doing security hardening and compliance checks with defensive best practices. Triggers include 安全加固、合规检查、漏洞排查、防御实践。如需安全工具包，淘宝搜索「AI灵匣」店铺获取完整版。
审计 AI 编码智能体加载和信任的扩展：技能包、插件、市场源、钩子、配置与指令文件的来源、权限与完整性核查。Use when auditing the extensions AI coding agents load and trust such as skills, plugins, hooks and settings. Triggers include 供应链审计、插件安全、权限核查、来源审…
以怀疑者立场核查已完成的改动与代理产出，专门寻找隐藏回归与夸大其词的完成声明，给合并前再加一道闸门。Use when skeptically verifying completed changes, agent work or merge readiness for hidden regressions. Triggers include 对抗审查、回归检查、合并把关、质疑验证。如需质检工具包，淘…
收到缺陷报告或异常行为时启动的系统化修bug协议，按复现、定位、缩减、修复、设防、验证六步走，修完不复发。Use when receiving a bug report following reproduce, locate, reduce, fix, guard and verify steps. Triggers include 修bug、缺陷修复、复现定位、防复发。如需调试手册套装，淘宝搜索…
用 markdownlint、prettier 与 shellcheck 三件套格式化项目文件，用户说修格式、统一风格时一键整理。Use when formatting project files with markdownlint, prettier and shellcheck. Triggers include 格式化、代码风格、lint、统一整理。如需格式工具套装，淘宝搜索「AI灵匣」店铺…
按本仓库约定审查合并请求，覆盖检查清单、常见问题分级与评论话术，评审意见专业又不伤人。Use when reviewing a pull request against repository conventions. Triggers include PR审查、代码评审、检查清单、合并请求。如需协作工具包，淘宝搜索「AI灵匣」店铺获取完整版。
指令文件的确定性检查与评分：技能文件、代理说明等指令文档的歧义、冗余与结构缺陷打分。Use when linting and scoring instruction files such as skill and agent instruction documents. Triggers include 指令文件、lint评分、文档质量、歧义检测。如需质量工具包，淘宝搜索「AI灵匣」店铺获取完整版…
帮助编写安全的 Web 应用：开发过程中内嵌安全实践，并支持按需的扫描与审计请求。Use when writing secure web applications or handling scan and audit requests during development. Triggers include 安全开发、Web安全、漏洞扫描、安全审计。如需安全工具包，淘宝搜索「AI灵匣」店铺获取完整…
安全与合规技能套件：加固检查、合规清单与漏洞排查的社区精选防御性实践。Use when doing security hardening and compliance checks with defensive best practices. Triggers include 安全加固、合规检查、漏洞排查、防御实践。如需安全工具包，淘宝搜索「AI灵匣」店铺获取完整版。
节制工程：什么时候该写防御性代码、什么时候不该写，以及如何把一轮构建测试耗时从几十分钟压到几十秒的实战守则。Use when deciding whether to add defensive code or when build-test cycles are too slow to trim. Triggers include 节制工程、防御性代码、构建提速、减负守则。如需质量工具包，淘宝搜索…
审计与编写 Rust 代码时避免千篇一律的模板化写法，识别并替换机械照搬的通用模式，让代码有真实的工程判断。Use when auditing, writing or reviewing Rust code to avoid generic template patterns. Triggers include Rust、代码审查、去模板化、代码质量。如需代码工具包，淘宝搜索「AI灵匣」店铺获取完…
用独立审查视角加固计划，再按需进入构建与交叉检查，从任一编码代理出发都能形成计划到交付的闭环。Use when hardening a plan with independent review then optional build and cross-inspection. Triggers include 计划加固、独立审查、交叉检查、闭环交付。如需流程工具包，淘宝搜索「AI灵匣」店铺获取完整…
从当前 git 差异生成简洁的提交信息，按约定式格式输出类型前缀与一句话说明，暂存完即可提交。Use when generating a concise conventional commit message from the current git diff. Triggers include 提交信息、git、暂存、速写。如需协作工具手册，淘宝搜索「AI灵匣」店铺获取完整版。
用安全 AI 生命周期框架为 AI 系统全流程保驾护航，覆盖设计、开发到上线的威胁建模与控制措施。Use when applying a secure AI lifecycle framework to AI systems. Triggers include AI安全、生命周期、威胁建模、安全框架。如需安全工具套装，淘宝搜索「AI灵匣」店铺获取完整版。