spacesafeml-certification-benchmark-and-/AGENTS.md

Architecture and contribution guide for SpaceSafeML MVP

Overview

• Language: Python
• Core layout (src/spacesafeml_certification_benchmark_and_):
  • init.py
  • dsl.py (# DSL data models for safety contracts)
  • verification.py (# Verification engine and certificate generation)
  • simulation/ (# Simulation templates and HIL scaffolding)
  • governance/ledger.py (# Tamper-evident ledger for test attestations)
  • adapters/ (# Starter adapters for planner and perception modules)

Workflow

• Tests: pytest-based tests run via test.sh in repo root
• Build: python -m build and packaging metadata validated by test.sh
• Docs: README and AGENTS.md describe architecture and contribution rules

Tech stack

• Python 3.8+
• Lightweight verification (model-checker heuristic) and property-based testing scaffold
• Simple deterministic simulation scaffolds for fleet scenarios
• Tamper-evident ledger using hash chaining

Testing commands

• Run tests: bash test.sh
• Run Python build: python3 -m build
• Run unit tests only: pytest -q

Coding guidelines

• Keep modules small and well-scoped
• Prefer minimal, well-documented interfaces
• Add tests with each feature, ensure green before publishing
• Do not alter license or packaging metadata without explicit approval

Contribution rules

• Open PRs should include updated tests and a brief changelog-like summary
• All code must pass tests locally