CosmosMesh Privacy-Preserving Federated Mission Planning

This repository contains a production-oriented MVP of CosmosMesh, a privacy-preserving, offline-first federation platform for distributed mission planning across diverse space assets (rovers, drones, habitat modules, and orbiting satellites).

Key concepts

• Local optimization problems (contracts) per asset with explicit data contracts and versioning.
• Federated optimization through lightweight sharing of summarized signals (primal/dual variables) and aggregated statistics.
• Global assembly of local problems into a feasible fleet plan with delta-sync and audit trails.
• Privacy-by-design: secure aggregation, optional local differential privacy, and role-based access to signals.
• Identity and security: DIDs, short-lived certificates, tamper-evident logging.
• Adapters and simulators to bootstrap cross-domain interoperability and testing.
• Open API and governance ledger for provenance and interoperability.

This repository provides a minimal yet production-minded surface to bootstrap the ecosystem, with a focus on small, well-scoped changes and testable interfaces.

How to run tests

• Ensure dependencies are installed via the packaging metadata in pyproject.toml.
• Run tests: bash test.sh
• Build the package: python3 -m build

Note: The MVP is deliberately lean. The roadmap includes a canonical EnergiBridge for vendor-agnostic interoperability, a minimal DSL, and two starter adapters to bootstrap cross-domain demonstrations.

For maintainers

• Follow the architecture described in AGENTS.md and keep changes small and well-documented.
• Add tests for any public surface changes.
• If new dependencies are introduced, update pyproject.toml accordingly.