AI Testing Agent Framework · Open-Source · Multi-LLM · One-command deploy
English | 简体中文
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 learnoutputs theory references at every step (22 KB cards across 13 domains) - Safe-by-default — sandboxed exec · PII scrub · prompt-injection scan · decisions audit trail
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 before running pentest or AI-adversarial workflows.
# Download install script
curl -fsSL -o install.py https://raw.githubusercontent.com/Wool-xing/Test-Agent/main/install.py
# Deploy to your project directory (any path works)
python install.py D:\Test-Agent # Windows example, any drive or folder
python install.py ~/test-agent-project # macOS / Linux example, any folderWindows users: if curl fails with
CRYPT_E_NO_REVOCATION_CHECK, use PowerShell:Invoke-WebRequest -Uri https://raw.githubusercontent.com/Wool-xing/Test-Agent/main/install.py -OutFile install.py python install.py D:\Test-Agent # example, any drive or folder works
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: cp config/.env.example .env → edit .env → cd project-dir && claude → read skills/smoke-test.md workflow, run agents per the flow
No Python / Node / Docker required. Download and run.
| Upload | Dashboard | Catalog | History |
|---|---|---|---|
 |
 |
 |
 |
| Audience | Read |
|---|---|
| First-time user | Quick start → Deploy |
| QA engineer | User manual → Skill catalog |
| Architect / SRE | Architecture deep-dive → Runtime modules |
| Security researcher | Pentest expert → pentest-coordinator |
| Automotive tester | Automotive expert → ASIL workflow |
| Contributor | CONTRIBUTING.md → Marketplace |
- 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)
Test-Agent/
├── agents/ ← 16 expert agents
├── skills/ ← 32 business skills + 3 meta-skills
├── utils/ ← 79 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
See FULL_GUIDE.md and CHANGELOG.md for details.
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
See CONTRIBUTING.md. Community marketplace passes 4-gate verification.
MIT License — see LICENSE. Upstream components retain their own licenses; see NOTICE.md.
Made for testers · Built with testers · Tested by testers