aispec-best-practices — independently scanned and version-tracked by SaferSkills.
SaferSkills independently audited aispec-best-practices (Agent Skill) and scored it 100/100 (green). The audit ran 55 deterministic rules across Security, Supply Chain, Maintenance, Transparency, and Community; it found 0 high-severity and 0 lower-severity findings. The full rule-by-rule trace and per-finding evidence are below. Free, methodology-open.
Findings & checks · 0 flagged
Every scanned point with the score it earned and what moved between them.
First recorded scan — no prior version to compare against.
The primary manifest — the file an agent reads to learn what this artifact does.
检测项目技术栈,将对应的最佳实践规约写入项目。无需脚本 — AI 读取本文件即可执行。
检查 package.json、pyproject.toml、Cargo.toml、pubspec.yaml、CMakeLists.txt、go.mod、pom.xml、build.gradle,以及 src-tauri/、qml/ 等目录。
空目录? 问用户 1-2 个问题:项目类型 + 具体框架,然后从下表中选择。
规约位于 best-practices/zh/(中文)和 best-practices/en/(英文)。根据用户消息语言选择。
| 技术栈 | 路径(best-practices/{lang}/ 下) |
|---|---|
| Next.js / React 全栈 | frontend/nextjs-vercel-reference.md |
| Vue / Nuxt 3 | frontend/vue-nuxt-reference.md |
| React Native / Expo | frontend/react-native-reference.md |
| Flutter | frontend/flutter-app-reference.md |
| 小程序(微信/Taro/uni-app) | miniapp/miniapp-reference.md |
| Tauri(Rust+React)桌面 | desktop/tauri-rust-react-desktop-reference.md |
| Qt(C++/QML)桌面 | desktop/qt-desktop-reference.md |
| Python Qt(PySide6/PyQt6) | desktop/python-qt-desktop-reference.md |
| Go 后端 | backend/go-project-reference.md |
| Java / Spring Boot | backend/java-backend-reference.md |
| Python / FastAPI | backend/python-backend-reference.md |
| Node.js / Fastify / NestJS | backend/nodejs-backend-reference.md |
| Rust 系统服务 | system-tools/rust-system-service-reference.md |
| CLI(Rust/Go/Python/Node) | system-tools/cli-tool-reference.md |
| 技能开发 | system-tools/skill-development-reference.md |
| AI Agent / LLM / RAG | ai-and-data/ai-agent-llm-reference.md |
| 数据 / ML 工程 | ai-and-data/data-ml-project-reference.md |
前端/桌面项目还可参考 frontend/ui-style-and-icons-reference.md。
读取规约文件,将 [项目名] 替换为实际项目名,然后写入:
| 编辑器 | 写入路径 | 说明 |
|---|---|---|
| Cursor | .cursor/rules/project-spec.mdc | 一条规则一个文件,含 frontmatter |
| Claude Code | .claude/CLAUDE.md 或 .claude/rules/project-spec.md | 追加或新建;主规约 + 模块化规则均可 |
| OpenClaw | 见下方 OpenClaw 说明 | 默认从 workspace 读,非项目根;可写项目根以随仓版本管理 |
| Codex | AGENTS.md 或 .codex/rules/ | 按 Codex 文档配置 |
| Antigravity (Google) | .antigravity/rules.md 或 .agent/rules/project-spec.md | 项目规则自动加载 |
| Trae(字节) | .trae/project_rules.md 或 .trae/rules/project-spec.md | Markdown 格式 |
| Code Buddy(腾讯云) | .codebuddy/rules/project-spec/RULE.mdc 或根目录 AGENTS.md | 每条规则一个子目录含 RULE.mdc |
若无法判断编辑器,默认按 Cursor 格式写入。
OpenClaw 说明:写代码时默认以 Claude Code 为执行方(如在 OpenClaw 中调起 Claude Code 写代码)。OpenClaw 从 workspace(配置项 agent.workspace,常见为 ~/.openclaw/workspace)读取 AGENTS.md、SOUL.md 等,不是从“当前项目根目录”读。本技能仍建议在当前项目根目录生成 AGENTS.md(及可选 SOUL.md),以便规约随仓库版本管理、团队共享;用户若希望 OpenClaw 使用该规约,可:(1) 将 OpenClaw 的 workspace 指向项目根,或 (2) 把项目根下的 AGENTS.md 复制/链接到自己的 workspace。规约文件都写在当前项目根,不写到 ~/.openclaw/ 或其它项目。
列出检测到的技术栈、选择的规约及输出路径。建议用户查看 best-practices/{lang}/README.md 了解更多可用规约。
best-practices/ 中的内容,不可自行编造。best-practices/(技能不在上下文中),提示用户克隆仓库或运行 aispec-skill/scripts/install.sh。~30 seconds. Free. No account. Every finding cites a rule and a line of evidence.