How wool-xing/test-agent works
Test-Agent is an AI-powered software testing framework: users submit a requirement or target through the CLI, API, or web UI; the runtime routes work to specialist agents and testing capabilities, then exposes run status and reports. The README also documents multi-platform/protocol testing, compliance and quality gates, learning mode, evidence, integrations, and a plugin marketplace. Some subsystem links below are documented capabilities rather than inspected call paths; those edges cite the README, and unsampled implementation details are not inferred.
Main entry points: direct.py Run Orchestrator · llm_client.py LLM Client · main.py Runtime API · base.py Specialist Agents
Components 21
User Interfaces
- CLI
runtime/cli/main.py - Runtime API
runtime/api/main.py - Web Console
runtime/web/src/App.tsx
Test Execution
- Run Orchestrator
runtime/orchestrator/direct.py - Specialist Agents
runtime/orchestrator/agents/base.py - Model Router
runtime/router/model_router.py - LLM Client
runtime/router/llm_client.py - Testing Capabilities
utils - Execution Sandbox
runtime/core/sandbox.py
Analysis and Quality
- Quality Gates
utils/quality/quality_gate_engine.py - Impact Analysis
runtime/intelligence/impact_engine.py - Compliance Checks
runtime/compliance/engine.py - Reports and Exports
runtime/exporters - Learning Mode
runtime/learning_loop
Integrations and Extensions
- Messaging Gateway
runtime/gateway - Plugin Marketplace
runtime/marketplace
Runtime Foundation
- Run Results
runtime/storage/models.py - Evidence Vault
runtime/mcp/evidence_vault/server.py - Runtime Settings
runtime/config/settings.py
Other components
- Test Engineer
- LLM Providers
Connections 21
- Test Engineer → CLI: submits requirements
README.md - Test Engineer → Runtime API: submits target
README.md - Test Engineer → Web Console: uses console
README.md - Runtime API → Run Orchestrator: submits run
runtime/api/main.py - Run Orchestrator → Specialist Agents: dispatches work
README.md - Specialist Agents → LLM Client: requests completion
runtime/orchestrator/agents/base.py - LLM Client → Model Router: resolves model
runtime/router/llm_client.py - LLM Client → LLM Providers: calls provider
runtime/router/llm_client.py - Run Orchestrator → Testing Capabilities: runs tests
README.md - Run Orchestrator → Execution Sandbox: executes safely
README.md - Run Orchestrator → Run Results: records run
README.md - Runtime API → Run Results: returns status
runtime/api/main.py - Run Orchestrator → Evidence Vault: captures evidence
README.md - Run Orchestrator → Reports and Exports: produces reports
README.md - Run Orchestrator → Quality Gates: evaluates gates
README.md - Run Orchestrator → Compliance Checks: checks compliance
README.md - Impact Analysis → Testing Capabilities: recommends tests
runtime/intelligence/impact_engine.py - Runtime Settings → LLM Client: configures providers
runtime/router/llm_client.py - Test Engineer → Learning Mode: enables learn mode
README.md - Run Orchestrator → Messaging Gateway: notifies channels
README.md - Test Engineer → Plugin Marketplace: discovers plugins
README.md