AidMesh: Federated, Privacy-Preserving Disaster Response Orchestrator

Overview

• AidMesh provides a privacy-preserving, offline-first orchestration framework for disaster relief planning across partner organizations. It models core primitives as a Graph-of-Contracts (GoC) with:
• Objects: LocalProblems (per-site relief tasks)
• Morphisms: SharedSignals and DualVariables
• PlanDelta: cryptographically-tagged actions
• AuditLog / PrivacyBudget: provenance and privacy controls
• Time rounds: delta-sync cadence for islanded operation and deterministic replay

Delivery Goal

• A production-ready skeleton with a small, testable core ensuring end-to-end delta-sync between partners over TLS-like channels, plus a toy ADMM-lite solver for cross-organization optimization.

What’s included

• Core data models (dataclasses)
• Lightweight ADMM-lite solver
• Toy adapters for onboarding (SupplyDepotController, FieldDistributionPlanner)
• Basic tests and packaging metadata (pyproject.toml)
• test.sh that runs pytest and builds the package

How to use

• Run tests: ./test.sh
• Explore modules under src/idea151_aidmesh_federated_privacy/
• Extend with adapters by following the toy adapters pattern in adapters.py

Contribution

• See AGENTS.md for architectural details and contribution guidelines.

Note: This repository is intended as a production-ready, testable chunk that can be extended by subsequent teams in the SWARM.