Wool-xing/Test-Agent

Test-Agent is an open-source AI testing agent framework that automates testing from requirement parsing to PoC-validated bug reports. It turns software (EXE, APK, Docker image, API) into a fully tested project using 16 expert AI agents and a skills catalog. Typical users: QA engineers, security researchers, SRE/architects, automotive testers and contributors. Key features: - Broad platform/protocol support: Web, API, Mobile, Desktop, IoT, Automotive, Blockchain; HTTP/gRPC/WebSocket/TCP/UDP/GraphQL/MQTT/CAN-bus/Kafka/Modbus etc. - Multi-LLM, no vendor lock-in: six built-ins (Claude, OpenAI, Gemini, Qwen, DeepSeek, Ollama) plus OpenAI-compatible fallback; configure via env vars. - One-command deploy (install.py) or desktop app; outputs test artifacts (Excel, xmind, markmap, OPML), Word reports and decision logs. - Safety and auditability: sandboxed execution, PII scrub, prompt-injection scan, decision trail. - Learn mode (--mode learn) provides theory references; supports many test types (functional, performance, security, compatibility, stability, accessibility, visual, AI-specific). Project contains agents, skills, utils, runtime and marketplace; tech stack includes pytest, Playwright, Appium, JMeter, FastAPI and Postgres+pgvector.

AI-related keywords found in this repo's description, topics, or README summary — grouped by category. Each badge links to the corresponding ranking detail page.