- PRD.md complete
- ARCHITECTURE.md complete
- ROADMAP.md complete
- METHODOLOGY.md complete
- Initial git commit
- Folder structure created
- pyproject.toml configured
- .gitignore configured
- Basic package imports working
- Needle generation (synthetic)
- Needle generation (template)
- Needle generation (relational/ATC)
- Haystack builder (Paul Graham source)
- Haystack builder (Wikipedia source)
- Depth injector
- Evaluator (exact match)
- Evaluator (semantic)
- Evaluator (LLM judge)
- Runner (async)
- Runner (checkpoint/resume)
- Runner (caching)
- Anthropic provider
- OpenAI provider
- Google provider
- Mistral provider
- Cohere provider
- Groq provider
- Ollama provider (local)
- S-RT task
- M-RT task
- M-RS task
- ATC task
- Counting task
- Key-Value task
- Conflicting needles task
- SQLite schema and migrations
- Credential storage (keyring)
- Response caching
- Heatmap generator
- Line chart generator
- Radar chart generator
- HTML report generator
- Setup commands
- Run commands
- Results commands
- Serve command
- FastAPI app skeleton
- Provider/credential routes
- Experiment routes
- Results routes
- WebSocket progress streaming
- Project scaffold (SvelteKit)
- Dashboard page
- Model setup page
- New experiment wizard
- Progress tracking page
- Results dashboard with heatmap
- Results library
- Unit tests (80%+ coverage)
- Integration tests
- README.md
- QUICKSTART.md
- CONTRIBUTING.md
- API docs (auto-generated)
- CI/CD pipeline
- Package builds successfully
- All docs reviewed
- Git tags and CHANGELOG