From d22b196aba4a02d2a2d21b7c26cf93124d5bb650 Mon Sep 17 00:00:00 2001 From: xiaoxing0135 <706015750@qq.com> Date: Tue, 2 Jun 2026 01:45:03 +0800 Subject: [PATCH] =?UTF-8?q?docs:=20restructure=20READMEs=20=E2=80=94=20mer?= =?UTF-8?q?ge=20intro+features,=20simplify=20structure,=20cut=20redundant?= =?UTF-8?q?=20sections?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- README.md | 127 ++++++++++++++++------------------------------ README.zh-CN.md | 131 +++++++++++++++++------------------------------- 2 files changed, 89 insertions(+), 169 deletions(-) diff --git a/README.md b/README.md index ab8470a..400b9ac 100644 --- a/README.md +++ b/README.md @@ -13,11 +13,21 @@ --- +Test-Agent turns any software, EXE, APK, Docker image, or API into a **fully tested project** — autonomous from requirement parsing to PoC-validated bug reports, powered by 16 AI agents. + +- **All-platform** — Web · API · Android · iOS · WeChat Mini Program · Windows · macOS · Linux · Automotive · IoT · AI/LLM · Blockchain +- **All-protocol** — HTTP(S) · gRPC · WebSocket · TCP/UDP · GraphQL · MQTT · CAN-bus · Kafka · Modbus · … +- **Multi-LLM no lock-in** — 6 built-in (Claude / OpenAI / Gemini / Qwen / DeepSeek / Ollama) + OpenAI-compatible fallback for any provider, 3 env vars, zero code +- **Learn while using** — `--mode learn` outputs theory references at every step (22 KB cards across 13 domains) +- **Safe-by-default** — sandboxed exec · PII scrub · prompt-injection scan · decisions audit trail + +--- + ## 🚀 Quick Start > Prerequisite: Python 3.x. Git & Node.js auto-installed if missing (winget / brew / apt / apk). > -> ⚠️ This project includes offensive security tools (pentest skill / SSRF probes / AI adversarial templates). Read [SECURITY.md](SECURITY.md) authorization requirements before running pentest or AI-adversarial workflows. +> ⚠️ This project includes offensive security tools (pentest skill / SSRF probes / AI adversarial templates). Read [SECURITY.md](SECURITY.md) before running pentest or AI-adversarial workflows. ```bash # Download install script @@ -34,40 +44,17 @@ python install.py ~/test-agent-project # macOS / Linux example, any folder > python install.py D:\Test-Agent # example, any drive or folder works > ``` -**Expected duration**: ~10–15 min (includes pip install + Playwright chromium download). CN networks auto-detect and use Tsinghua PyPI mirror. +**Expected duration**: ~10–15 min (includes pip deps + Playwright browser). CN networks auto-use Tsinghua PyPI mirror. After deployment, outputs under `workspace/`: test cases (Excel + xmind + markmap + opml) + Word report + decision logs. -**Next**: edit `.env` → `claude /login` → `cd project-dir && claude` → inside Claude: `/smoke-test` - ---- - -## What is Test-Agent? - -Test-Agent turns any software, EXE, APK, Docker image, or API into a **fully tested project** — autonomous from requirement parsing to PoC-validated bug reports. Built for QA teams, security researchers, automotive testers, and anyone who wants to **use AI testing while learning the theory behind it**. +**Next**: edit `.env` → `claude /login` → `cd project-dir && claude` → `/smoke-test` -- **16 expert agents** (11 production + 5 script — V1.x rollout 收尾,见 [ROADMAP.md](ROADMAP.md)) — functional · security · mobile · desktop · AI model · automotive · pentest … -- **30 active skills** (23 production + 7 script) **+ 2 vision-only** (reference, not executable) **+ 3 meta-skills** — TDD · E2E · regression · pentest · car-CAN-bus · eval-harness · … -- **78 production utils** — pytest · Playwright · JMeter · Appium · Burp · Allure · OpenCV · … -- **Multi-LLM (any provider, plug-and-play)** — 6 built-in (Claude / OpenAI / Gemini / Qwen / DeepSeek / Ollama) + **OpenAI-compatible fallback channel** for any other provider (Zhipu / Doubao / Kimi / Baichuan / Xunfei / …) via 3 env vars, zero code change. Cookbook: [`config/llm-providers.md`](config/llm-providers.md) -- **BugTracker** — 1 active adapter (Zentao); 5 planned (Jira · GitHub · GitLab · Linear · Webhook, see roadmap) -- **6 notify channels** — WeChat Work · Lark/Feishu · DingTalk · Slack · Email · MS Teams -- **MCP integration** — 6 server modules implemented (test-orchestrator active by default; 5 others ready to enable in `.mcp.json`) -- **Self-test scaffolding** — L1 lint + L2 mock CI active in CI; L3 real-LLM + L4 weekly cron require `ANTHROPIC_API_KEY` secret (not configured in this repo by default) - -## 🖥 Desktop App (Windows + macOS) +## 🖥 Desktop App [![Download](https://img.shields.io/badge/Download-Latest%20Release-blue)](https://github.com/Wool-xing/Test-Agent/releases) -No Python/Node/Docker required. Download the installer for your platform and start testing immediately. - -## 🎯 5 Key Capabilities - -1. **All-platform** — Web / API / Android / iOS / WeChat-miniprogram / Windows EXE / macOS / Linux / Electron / game / IoT / audio-video / AI/LLM / blockchain / 车载 -2. **All-protocol** — HTTP(S) / gRPC / WebSocket / TCP / UDP / GraphQL / SOAP / MQTT / SSH / serial / Kafka / RabbitMQ / Modbus / CAN-bus / SOME-IP / DoIP / UDS -3. **Multi-LLM no lock-in (any provider)** — 6 built-in providers via `tagent config use ` (Claude / OpenAI / Gemini / Qwen / DeepSeek / Ollama) plus **OpenAI-compatible fallback** via `tagent config use-compat` for any other (Zhipu / Doubao / Kimi / Baichuan / Xunfei / …) — 3 env vars, zero code change. See [`config/llm-providers.md`](config/llm-providers.md) -4. **Learn while using** — `--mode learn` outputs every step with theory references (22 KB cards across 13 domains: tools / coding / foundation / strategy / methods / protocols / platforms / gates / security / AI testing / compliance / process / build-your-own) -5. **Safe-by-default** — sandboxed exec / PII scrub / runtime prompt-injection scan / 4-gate marketplace verify / decisions audit trail +No Python / Node / Docker required. Download and run. ## 📸 Screenshots @@ -75,82 +62,52 @@ No Python/Node/Docker required. Download the installer for your platform and sta |--------|-----------|---------|---------| | ![Upload](docs/assets/screenshots/upload.png) | ![Dashboard](docs/assets/screenshots/dashboard.png) | ![Catalog](docs/assets/screenshots/catalog.png) | ![History](docs/assets/screenshots/history.png) | -## 📊 Coverage - -- **Product types**: Web · API · Mobile · Desktop · IoT · AI · Blockchain · Vehicle · Embedded · Serverless -- **Test types**: functional / performance / security / compatibility / weak-network / stability / reliability / accessibility / contract / visual / i18n / observability / chaos / mutation / AI-specific (hallucination / prompt-injection / drift / fairness) / compliance -- **Test design methods**: equivalence-partitioning · boundary-value · decision-table · state-transition · pairwise · orthogonal · exploratory SBTM · risk-based · TDD · BDD · ATDD -- **Quality gates**: smoke → regression → performance_ci_quick → performance_full → release (5-layer) - -Coverage across the listed categories is broad but not exhaustive. Domain-specific gates (DO-178C avionics / HIPAA medical / IEC 61508 industrial) ship as skeleton compliance YAML profiles — production use in regulated industries requires domain expert review. - -## 📖 Design Documents - -For project design rationale, architecture decisions, and methodology rationale, see [FULL_GUIDE.md](FULL_GUIDE.md). Inspirations from upstream OSS (hermes / gbrain / karpathy / etc.) are credited in [NOTICE.md](NOTICE.md). - -## 📂 Project Structure - -```text -Test-Agent/ -├── 00-项目导航.md ← 5-dimension category guide -├── docs/getting-started/ ← user manual / deploy / config / deliverables -├── agents/ ← 16 expert agents (11 production + 5 script, V1.x rollout 收尾) -├── skills/ ← 32 business skills (23 production + 7 script + 0 rollout + 2 vision) + 3 meta-skills -├── config/ ← conftest / pytest.ini / .env / .mcp.json -├── utils/ ← 78 production utils -├── ci/ ← GitHub Actions + Jenkins -├── runtime/ ← V1.x runtime layer (router / orchestrator / MCP / web / scheduler / subagent / learning_loop / backends / gateway / tutor / essence_watcher / marketplace) -├── docs/charter/ ← Vision charter (7 split files: vision-dimensions / coverage-matrix / agentchat-protocol / skills-bugtracker / install-deploy / test-architecture / runtime-license) -├── docs/theory/ ← 22 teaching KB cards across 13 categories -├── profiles/compliance/ ← 10 industry compliance YAML profiles -├── marketplace/ ← Community skills / agents / mcp / hooks (4 lanes, 4-gate verify) -├── install.py ← one-command cross-platform deploy -├── README.md ← This file -├── FULL_GUIDE.md ← Full engineering guide -├── CHANGELOG.md ← Version log -└── LICENSE / SECURITY.md / CONTRIBUTING.md / CODE_OF_CONDUCT.md -``` - -> **Skill Lifecycle (meta-tools)**: -> - **Current (A · methodology reference)**: Each subdir's SKILL.md serves as skill-design reference material. -> - **Usable today (B · perspective extension)**: Use `nuwa-skill` to distill new mental-model perspectives (Naval / Munger / Feynman); use `darwin-skill` to optimize perspective skills. -> - **V2.x Roadmap (C · testing-domain adaptation)**: Re-target nuwa as a test skill/agent distiller; re-target darwin's 8-dim scoring to testing domain. - ## 📚 Documentation | Audience | Read | |----------|------| | **First-time user** | [Quick start](docs/getting-started/INDEX.md) → [Deploy](docs/getting-started/部署说明.md) | | **QA engineer** | [User manual](docs/getting-started/使用手册.md) → [Skill catalog](skills/) | -| **Architect / SRE** | [Architecture deep-dive](docs/charter/06-test-architecture.md) → [Runtime](docs/charter/07-runtime-license.md) → [Runtime modules](runtime/INDEX.md) | +| **Architect / SRE** | [Architecture deep-dive](docs/charter/06-test-architecture.md) → [Runtime modules](runtime/INDEX.md) | | **Security researcher** | [Pentest expert](agents/15-渗透测试.md) → [pentest-coordinator](skills/pentest-coordinator.md) | | **Automotive tester** | [Automotive expert](agents/16-车载测试.md) → [ASIL workflow](skills/automotive-test.md) | | **Contributor** | [CONTRIBUTING.md](CONTRIBUTING.md) → [Marketplace](marketplace/INDEX.md) | -## 🛠️ Tech Stack +## 📊 Coverage -pytest 8.3 · Playwright 1.59 · Appium 5.3 · pywinauto · JMeter 5.6 · Allure · Airtest · OpenCV · Faker · SQLAlchemy 2.0 · MCP 1.0 · LiteLLM · Prefect · FastAPI · React 18 · Tailwind · Postgres+pgvector · MinIO · OpenTelemetry · Loguru · Docker Compose · GitHub Actions / Jenkins +- **Product types**: Web · API · Mobile · Desktop · IoT · AI · Blockchain · Automotive · Embedded · Serverless +- **Test types**: functional / performance / security / compatibility / weak-network / stability / accessibility / visual / i18n / chaos / mutation / AI-specific / compliance +- **Quality gates**: smoke → regression → performance_ci_quick → performance_full → release (5-layer) -## 🤝 Contributing +## 📂 Project Structure -See [`CONTRIBUTING.md`](CONTRIBUTING.md) for the contribution workflow. +```text +Test-Agent/ +├── agents/ ← 16 expert agents +├── skills/ ← 32 business skills + 3 meta-skills +├── utils/ ← 78 production utils (pytest · Playwright · JMeter · Appium · …) +├── config/ ← conftest / pytest.ini / .mcp.json +├── runtime/ ← V1.x runtime (router · orchestrator · MCP · …) +├── ci/ ← GitHub Actions + Jenkins +├── docs/ ← user manual / architecture / theory / compliance +├── marketplace/ ← community skills / agents / mcp / hooks +├── install.py ← one-command cross-platform deploy +└── README.md / README.zh-CN.md +``` -Community marketplace contributions (`marketplace/`) pass through **4 verification gates** (current implementation): signature presence (planned) → injection-regex scan → AST syntax-parse (V1.x: replace with real Docker sandbox) → frontmatter-presence score (V1.x: swap for real darwin-skill evaluator). +See [FULL_GUIDE.md](FULL_GUIDE.md) and [CHANGELOG.md](CHANGELOG.md) for details. -## 📜 License +## 🛠️ Tech Stack + +pytest 8.3 · Playwright 1.59 · Appium 5.3 · JMeter 5.6 · Allure · OpenCV · SQLAlchemy 2.0 · MCP 1.0 · LiteLLM · FastAPI · React 18 · Postgres+pgvector · Docker -MIT License — see [LICENSE](LICENSE). +## 🤝 Contributing -Upstream components retain their own licenses; see [NOTICE.md](NOTICE.md) for attributions. +See [CONTRIBUTING.md](CONTRIBUTING.md). Community marketplace passes 4-gate verification. -## 🙏 Inspirations (essence absorbed) +## 📜 License -- [hermes-agent](https://github.com/NousResearch/hermes-agent) — closed learning loop + 7 backends + multi-platform gateway -- [gbrain](https://github.com/garrytan/gbrain) — self-wiring KB graph + eval replay + safe-by-default -- [andrej-karpathy-skills](https://github.com/forrestchang/andrej-karpathy-skills) — 4 LLM-coding principles -- [everything-claude-code](https://github.com/affaan-m/everything-claude-code) — TDD / verification / harness-first -- [pentagi](https://github.com/vxcontrol/pentagi) + [shannon](https://github.com/KeygraphHQ/shannon) — pentest agent black-box + white-box -- [build-your-own-x](https://github.com/codecrafters-io/build-your-own-x) — deep-dive learning path +MIT License — see [LICENSE](LICENSE). Upstream components retain their own licenses; see [NOTICE.md](NOTICE.md). --- diff --git a/README.zh-CN.md b/README.zh-CN.md index 275b06c..ed5c96f 100644 --- a/README.zh-CN.md +++ b/README.zh-CN.md @@ -1,6 +1,6 @@ # 🤖 Test-Agent -> **AI 测试 Agent 框架 · 开源 · 多 LLM · 5 秒上手** +> **AI 测试 Agent 框架 · 开源 · 多 LLM · 一键部署** [![CI](https://github.com/Wool-xing/Test-Agent/actions/workflows/ci.yml/badge.svg?branch=main)](https://github.com/Wool-xing/Test-Agent/actions/workflows/ci.yml) [![Python](https://img.shields.io/badge/Python-3.11+-blue.svg)](https://www.python.org) @@ -13,11 +13,21 @@ --- +Test-Agent 将任何软件 / EXE / APK / Docker 镜像 / API 变成**完整测试过的项目**——从需求解析到 PoC 验证的 Bug 报告，由 16 个 AI Agent 全自主完成。 + +- **全平台** — Web · API · Android · iOS · 微信小程序 · Windows · macOS · Linux · 车载 · IoT · AI/LLM · 区块链 +- **全协议** — HTTP(S) · gRPC · WebSocket · TCP/UDP · GraphQL · MQTT · CAN-bus · Kafka · Modbus · … +- **多 LLM 无锁定** — 内置 6 厂商（Claude / OpenAI / Gemini / Qwen / DeepSeek / Ollama）+ OpenAI 兼容通道接任意厂商，3 个 env 零代码 +- **边用边学** — `--mode learn` 每步含理论引用（22 卡跨 13 大类） +- **safe-by-default** — 沙箱执行 · PII 脱敏 · Prompt 注入扫描 · decisions 审计链 + +--- + ## 🚀 快速开始 > 前提：Python 3.x。Git / Node.js 缺失时自动安装（winget / brew / apt / apk）。 > -> ⚠️ 本项目含攻击面工具（渗透 skill / SSRF 探针 / AI 对抗模板）。运行 pentest 或 AI-adversarial 工作流前请阅 [SECURITY.md](SECURITY.md) 中的授权要求。 +> ⚠️ 本项目含攻击面工具（渗透 skill / SSRF 探针 / AI 对抗模板）。运行 pentest 或 AI-adversarial 工作流前请阅 [SECURITY.md](SECURITY.md)。 ```bash # 下载部署脚本 @@ -34,81 +44,23 @@ python install.py ~/test-agent-project # macOS / Linux 示例，可改为 > python install.py D:\Test-Agent # 示例，可改为其他盘符或目录 > ``` -**预期耗时**：约 10–15 min（含 pip install + Playwright chromium 下载）。CN 网络自动配清华 PyPI 镜像。 +**预期耗时**：约 10–15 min（含 pip 依赖 + Playwright 浏览器）。CN 网络自动配清华 PyPI 镜像。 部署后 `workspace/` 下即见产物：测试用例（Excel + xmind + markmap + opml）+ Word 报告 + 决策日志。 -**下一步**：编辑 `.env` → `claude /login` → `cd 项目目录 && claude` → 在 Claude 内 `/smoke-test` - ---- - -## Test-Agent 是什么? - -Test-Agent 让任何软件 / EXE / APK / Docker 镜像 / API,变成**完整测试过的项目**——从需求解析到 PoC 验证的 Bug 报告,全自主。为 QA 团队、安全研究员、车载测试工程师、以及任何想**用 AI 测试同时学测试理论**的人而生。 - -- **16 专家 Agent** (11 production + 5 script — V1.x rollout 收尾,见 [ROADMAP.md](ROADMAP.md)) — 功能 · 安全 · 移动 · 桌面 · AI 模型 · 车载 · 渗透 …… -- **30 个活跃 Skill** (23 production + 7 script) **+ 2 个 vision-only**（方法论参考，不可执行）**+ 3 元 Skill** — TDD · E2E · 回归 · 渗透 · 车载 CAN · eval-harness …… -- **49 生产工具** — pytest · Playwright · JMeter · Appium · Burp · Allure · OpenCV …… -- **多 LLM(任厂商即插即用)** — 内置 6 厂商 (Claude / OpenAI / Gemini / Qwen / DeepSeek / Ollama) + **OpenAI 兼容兜底通道** 接其他任意厂商 (智谱 / 豆包 / Kimi / 百川 / 讯飞 / …), 3 个 env 变量, 零代码改动. 速查手册: [`config/llm-providers.md`](config/llm-providers.md) -- **BugTracker** — 1 已实装(禅道);5 计划(Jira · GitHub · GitLab · Linear · Webhook,见 roadmap) -- **6 通知渠道** — 企微 · 飞书 · 钉钉 · Slack · 邮件 · Teams -- **MCP 集成** — 6 模块已实现(test-orchestrator 默认启用;其余 5 件套在 `.mcp.json` 中按需启用) -- **自检脚手架** — L1 lint + L2 mock CI 已在 CI 激活;L3 真 LLM + L4 周自检需配置 `ANTHROPIC_API_KEY` secret(本仓默认未配) +**下一步**：编辑 `.env` → `claude /login` → `cd 项目目录 && claude` → `/smoke-test` -## 🖥 桌面应用 (Windows + macOS) +## 🖥 桌面应用 [![Download](https://img.shields.io/badge/下载-最新版本-blue)](https://github.com/Wool-xing/Test-Agent/releases) -无需安装 Python/Node/Docker。下载安装包即可开始测试。 - -## 🎯 5 大核心能力 - -1. **全平台** — Web / API / Android / iOS / 微信小程序 / Windows EXE / macOS / Linux / Electron / 游戏 / IoT / 音视频 / AI/LLM / 区块链 / 车载 -2. **全协议** — HTTP(S) / gRPC / WebSocket / TCP / UDP / GraphQL / SOAP / MQTT / SSH / 串口 / Kafka / RabbitMQ / Modbus / CAN-bus / SOME-IP / DoIP / UDS -3. **多 LLM 无锁定(任厂商)** — `tagent config use ` 切换 6 内置 (Claude / OpenAI / Gemini / Qwen / DeepSeek / Ollama), `tagent config use-compat` 接 **OpenAI 兼容兜底** 任意厂商 (智谱 / 豆包 / Kimi / 百川 / 讯飞 / …), 3 个 env 零代码. 见 [`config/llm-providers.md`](config/llm-providers.md) -4. **边用边学** — `--mode learn` 每步输出含**理论引用**(22 卡跨 13 大类:工具 / 编程 / 基础理论 / 策略 / 方法 / 协议 / 平台 / 门禁 / 安全 / AI 测试 / 合规 / 流程 / Build-Your-Own) -5. **safe-by-default** — 沙箱 / PII 脱敏 / 运行时 Prompt 注入扫描 / 4 关 Marketplace 验证 / decisions 审计链 - -## 📊 覆盖度 - -- **产品形态**:Web · API · 移动 · 桌面 · IoT · AI · 区块链 · 车载 · 嵌入式 · Serverless -- **测试类型**:功能 / 性能 / 安全 / 兼容 / 弱网 / 稳定 / 可靠性 / 可访问性 / 契约 / 视觉 / i18n / 可观测性 / 混沌 / 变异 / AI 特有(幻觉 / Prompt 注入 / 漂移 / 公平性)/ 合规 -- **用例设计方法**:等价类 · 边界值 · 判定表 · 状态迁移 · 配对 · 正交 · 探索性 SBTM · 风险驱动 · TDD · BDD · ATDD -- **质量门禁**:冒烟 → 回归 → performance_ci_quick → performance_full → release(5 层) - -覆盖面在上述类别较广但不穷尽。领域专项门禁（航空 DO-178C / 医疗 HIPAA / 工业 IEC 61508）以 skeleton 合规 YAML 形态提供，受监管行业生产使用需领域专家评审。 - -## 📖 设计文档 - -项目设计思路、架构决策、方法论详见 [FULL_GUIDE.md](FULL_GUIDE.md)。上游开源致谢见 [NOTICE.md](NOTICE.md)。 +无需 Python / Node / Docker，下载安装包即用。 -## 📂 项目结构 - -```text -Test-Agent/ -├── 00-项目导航.md ← 5 维度分类速查 -├── docs/getting-started/ ← 使用手册 / 部署 / 配置 / 交付物 -├── agents/ ← 16 个专家 Agent (11 production + 5 script + 0 rollout) -├── skills/ ← 32 个业务 Skill (23 production + 7 script + 0 rollout + 2 vision) + 3 个元 Skill -├── config/ ← conftest / pytest.ini / .env / .mcp.json -├── utils/ ← 78 个生产工具 -├── ci/ ← GitHub Actions + Jenkins -├── runtime/ ← V1.x 运行时(router/orchestrator/MCP/web/scheduler/subagent/learning_loop/backends/gateway/tutor/essence_watcher/marketplace) -├── docs/charter/ ← 愿景宪章(7 子文件: vision-dimensions / coverage-matrix / agentchat-protocol / skills-bugtracker / install-deploy / test-architecture / runtime-license) -├── docs/theory/ ← 22 教学 KB 卡片跨 13 大类 -├── profiles/compliance/ ← 10 行业合规 YAML -├── marketplace/ ← 社区 skills / agents / mcp / hooks(4 lane,4 关验证) -├── install.py ← 跨平台一键部署 -├── README.md / README.zh-CN.md -├── FULL_GUIDE.md ← 完整工程指南 -├── CHANGELOG.md ← 版本日志 -└── LICENSE / SECURITY.md / CONTRIBUTING.md / CODE_OF_CONDUCT.md -``` +## 📸 截图 -> **Skill 全生命周期(元工具)**: -> - **现状(A · 方法论参考)**:各子目录 SKILL.md 为 skill 设计参考材料。 -> - **当下可用(B · 人物视角扩展)**:用 `nuwa-skill` 蒸馏新人物视角(Naval / 芒格 / 费曼);用 `darwin-skill` 优化人物视角 skill。 -> - **V2.x 路线图(C · 测试领域适配)**:改造 nuwa 为测试 skill / agent 蒸馏器;改造 darwin 为测试领域 8 维评分。 +| 上传 | 仪表盘 | 目录 | 历史 | +|------|--------|------|------| +| ![上传](docs/assets/screenshots/upload.png) | ![仪表盘](docs/assets/screenshots/dashboard.png) | ![目录](docs/assets/screenshots/catalog.png) | ![历史](docs/assets/screenshots/history.png) | ## 📚 文档导航 @@ -116,35 +68,46 @@ Test-Agent/ |------|------| | **首次用户** | [快速开始](docs/getting-started/INDEX.md) → [部署说明](docs/getting-started/部署说明.md) | | **QA 工程师** | [使用手册](docs/getting-started/使用手册.md) → [Skill 目录](skills/) | -| **架构师 / SRE** | [架构深度](docs/charter/06-test-architecture.md) → [Runtime 章节](docs/charter/07-runtime-license.md) → [Runtime 模块](runtime/INDEX.md) | +| **架构师 / SRE** | [架构深度](docs/charter/06-test-architecture.md) → [Runtime 模块](runtime/INDEX.md) | | **安全研究员** | [渗透专家](agents/15-渗透测试.md) → [pentest-coordinator](skills/pentest-coordinator.md) | | **车载测试** | [车载专家](agents/16-车载测试.md) → [ASIL 工作流](skills/automotive-test.md) | | **贡献者** | [CONTRIBUTING.md](CONTRIBUTING.md) → [Marketplace](marketplace/INDEX.md) | -## 🛠️ 技术栈 +## 📊 覆盖度 -pytest 8.3 · Playwright 1.59 · Appium 5.3 · pywinauto · JMeter 5.6 · Allure · Airtest · OpenCV · Faker · SQLAlchemy 2.0 · MCP 1.0 · LiteLLM · Prefect · FastAPI · React 18 · Tailwind · Postgres+pgvector · MinIO · OpenTelemetry · Loguru · Docker Compose · GitHub Actions / Jenkins +- **产品形态**: Web · API · 移动 · 桌面 · IoT · AI · 区块链 · 车载 · 嵌入式 · Serverless +- **测试类型**: 功能 / 性能 / 安全 / 兼容 / 弱网 / 稳定性 / 可访问性 / 视觉 / i18n / 混沌 / 变异 / AI 特有 / 合规 +- **质量门禁**: 冒烟 → 回归 → performance_ci_quick → performance_full → release（5 层） -## 🤝 贡献 +## 📂 项目结构 -详见 [`CONTRIBUTING.md`](CONTRIBUTING.md)。 +```text +Test-Agent/ +├── agents/ ← 16 个专家 Agent +├── skills/ ← 32 个业务 Skill + 3 元 Skill +├── utils/ ← 78 个生产工具（pytest · Playwright · JMeter · Appium · …） +├── config/ ← conftest / pytest.ini / .mcp.json +├── runtime/ ← V1.x 运行时（router · orchestrator · MCP · …） +├── ci/ ← GitHub Actions + Jenkins +├── docs/ ← 使用手册 / 架构 / 教学 / 合规 +├── marketplace/ ← 社区 skills / agents / mcp / hooks +├── install.py ← 跨平台一键部署 +└── README.md / README.zh-CN.md +``` -社区 marketplace 贡献(`marketplace/`)走 **4 验证关**(当前实现):签名存在性检查(计划中)→ 注入正则扫 → AST 语法解析(V1.x:替换为真 Docker 沙箱)→ frontmatter 检测评分(V1.x:替换为真 darwin-skill 评估器)。 +详见 [FULL_GUIDE.md](FULL_GUIDE.md) 和 [CHANGELOG.md](CHANGELOG.md)。 -## 📜 许可 +## 🛠️ 技术栈 + +pytest 8.3 · Playwright 1.59 · Appium 5.3 · JMeter 5.6 · Allure · OpenCV · SQLAlchemy 2.0 · MCP 1.0 · LiteLLM · FastAPI · React 18 · Postgres+pgvector · Docker -MIT License — 详见 [LICENSE](LICENSE)。 +## 🤝 贡献 -上游组件保留各自协议;详见 [NOTICE.md](NOTICE.md) 致谢。 +详见 [CONTRIBUTING.md](CONTRIBUTING.md)。社区 marketplace 走 4 关验证。 -## 🙏 灵感来源(精髓汲取) +## 📜 许可 -- [hermes-agent](https://github.com/NousResearch/hermes-agent) — 封闭学习循环 + 7 后端 + 多平台 gateway -- [gbrain](https://github.com/garrytan/gbrain) — KB 自连图谱 + eval 回放 + safe-by-default -- [andrej-karpathy-skills](https://github.com/forrestchang/andrej-karpathy-skills) — LLM 写代码 4 原则 -- [everything-claude-code](https://github.com/affaan-m/everything-claude-code) — TDD / verification / harness-first -- [pentagi](https://github.com/vxcontrol/pentagi) + [shannon](https://github.com/KeygraphHQ/shannon) — 渗透 agent 黑盒+白盒 -- [build-your-own-x](https://github.com/codecrafters-io/build-your-own-x) — 深度学习路径 +MIT License — 详见 [LICENSE](LICENSE)。上游组件保留各自协议，详见 [NOTICE.md](NOTICE.md)。 ---