How xiayiyang831xyy/aitc-studio works
AITC is an AI test-engineering platform that turns requirements into reviewed, executable test assets. The main path runs from project and requirement input through feature confirmation, hybrid knowledge retrieval, skill-based generation, automated quality scoring, human approval, test execution, and evaluation. The React client also exposes knowledge, settings, and an Agent assistant; the backend includes persistent run governance and configurable model adapters. Workflow and capability links are based on README statements where implementation excerpts were unavailable.
Main entry points: registry.py Skill registry · graph.py Generation workflow · App.jsx AITC web app · api.js API client
Components 32
Web interface
- AITC web app
web/src/App.jsx - API client
web/src/services/api.js - Generation review
web/src/pages/GenerateFlow.jsx - Agent assistant
web/src/pages/AgentAssistant.jsx - Knowledge panels
web/src/pages/Knowledge.jsx - Test tasks
web/src/pages/ProjectTestTasks.jsx
Test production
- HTTP API
backend/app/main.py - Requirement intake
backend/app/api/requirements.py - Generation service
backend/app/services/generation_service.py - Generation workflow
backend/app/workflows/generation/graph.py - Quality checks
backend/app/services/quality_checker.py - Cases and execution
backend/app/models/testcase.py - Offline evaluation
backend/app/services/evaluation_service.py
AI capabilities
- Hybrid knowledge RAG
backend/app/services/knowledge_service.py - Skill registry
backend/app/skills/registry.py - Structured AI chains
backend/app/ai/chains.py - Model service
backend/app/services/llm.py - Endpoint security
backend/app/services/model_endpoint_security.py - Model settings
backend/app/services/settings_service.py - Prompt loader
backend/app/skills/shared/prompt_loader.py - Skill contracts
backend/app/skills/base.py - Mock skill support
backend/app/skills/shared/mock.py - Prompt formatting
backend/app/skills/shared/prompt_formatter.py
Run governance
- Runtime harness
backend/app/agent_runtime/harness.py - Run repository
backend/app/agent_runtime/repository.py - Test assistant agent
backend/app/agent/runner.py - Background worker
backend/app/worker.py
Persistence
- SQL database
backend/app/database.py - Agent run records
backend/app/models/agent_run.py - Generation records
backend/app/models/generation.py - Application config
backend/app/config.py
Other components
- Test engineer
Connections 39
- Test engineer → AITC web app: uses
README.md - AITC web app → Generation review: routes to
web/src/App.jsx - AITC web app → Agent assistant: routes to
web/src/App.jsx - AITC web app → Knowledge panels: routes to
web/src/App.jsx - AITC web app → Test tasks: routes to
web/src/App.jsx - Generation review → API client: requests
- Agent assistant → API client: requests
- Knowledge panels → API client: requests
- Test tasks → API client: requests
web/src/pages/ProjectTestTasks.jsx - API client → HTTP API: calls
- HTTP API → Requirement intake: mounts routes
backend/app/main.py - HTTP API → Generation service: mounts routes
backend/app/main.py - Requirement intake → Generation workflow: starts generation
- Generation service → Generation workflow: runs
- Generation workflow → Hybrid knowledge RAG: retrieves context
README.md - Generation workflow → Skill registry: dispatches skills
README.md - Generation workflow → Quality checks: checks candidates
README.md - Skill registry → Structured AI chains: invokes
- Structured AI chains → Model service: uses model
backend/app/ai/chains.py - Model service → Endpoint security: checks endpoints
backend/app/services/llm.py - Model service → Model settings: loads config
backend/app/services/llm.py - Skill registry → Skill contracts: uses contracts
backend/app/skills/registry.py - Skill registry → Prompt loader: loads prompts
backend/app/skills/shared/prompt_loader.py - Skill registry → Mock skill support: supports mock
- Skill registry → Prompt formatting: formats prompts
- Quality checks → Generation records: scores drafts
README.md - Generation service → Generation records: persists candidates
README.md - Generation review → Cases and execution: approves cases
README.md - Test tasks → Cases and execution: selects cases
README.md - Cases and execution → Offline evaluation: feeds evaluation
README.md - Agent assistant → Test assistant agent: drives tasks
README.md - Test assistant agent → Generation workflow: starts pipeline
README.md - Background worker → Run repository: claims runs
- Runtime harness → Run repository: records events
- Run repository → Agent run records: stores runs
backend/app/agent_runtime/repository.py - SQL database → Agent run records: creates tables
backend/app/database.py - SQL database → Generation records: creates tables
backend/app/database.py - SQL database → Application config: reads config
backend/app/database.py - Offline evaluation → Runtime harness: uses runtime
backend/app/services/evaluation_service.py