Skip to content

CONTRIBUTING.md

Latest commit

 

History

History
67 lines (45 loc) · 1.84 KB

CONTRIBUTING.md

File metadata and controls

67 lines (45 loc) · 1.84 KB
Raw

Contributing to LATTICE

Getting Started

git clone https://github.com/thunderstornX/lattice.git
cd lattice
pip install -e ".[dev]"
pytest

All 110 tests should pass before you start.

Code Standards

  • Type hints on every function
  • Docstrings on every public function and class
  • Functions under 20 lines (split if longer)
  • Specific exceptions (never bare except:)
  • No magic numbers
  • Python 3.10+ required

Running Tests

pytest tests/ -v

Running Benchmarks

PYTHONPATH=. python3 benchmarks/run_benchmarks.py

What We Need

Priority: Framework Adapters

Adapter plugins that hook into LangGraph, CrewAI, or AutoGen callback/event systems and automatically generate LATTICE claims. The ideal adapter requires zero changes to existing agent code.

Priority: Bayesian Confidence Propagation

The current effective confidence uses min-propagation (worst-case correct under correlated sources). Pluggable modules for alternative propagation strategies (Bayesian, weighted combination) behind a clean interface would be valuable for different threat models.

Priority: Scalability

Revocation waterfall and cycle detection show super-linear scaling past 1,000 claims. Pre-computed dependency indices or incremental transitive closure maintenance would help. See the paper (Section 6.3) for details.

Always Welcome

  • Bug reports with reproducible examples
  • Documentation improvements
  • Performance work (especially for large DAGs)
  • Test coverage improvements
  • Real-world case studies using LATTICE

Pull Request Process

  1. Fork the repo
  2. Create a branch from master
  3. Write tests for new functionality
  4. Make sure all tests pass
  5. Open a PR with a clear description of what changed and why

License

By contributing, you agree that your contributions will be licensed under the MIT License.