gridguard-secure-attested-cross-domain-e/AGENTS.md

GridGuard Agents and Architecture

Overview

• GridGuard uses a modular agent-based architecture to enable secure, attested cross-domain optimization.
• Agents (represented in code as AttestedAgent) are hardware-backed with remote attestation scaffolding and per-agent credentials tied to access contracts.
• The system orchestrates across modules: Security Contracts Registry, Verifiable Optimization, Transport, Governance Ledger, Delta-Sync, Adapters Marketplace, and a Simulation Harness.

Tech Stack (Python, MVP):

• Core language: Python 3.8+
• Dependencies are minimal; tests use pytest.
• Crypto-like primitives are simulated for MVP: signatures via SHA-256 digests; ZK-proof stubs emulate verification pipelines.

Testing and Verification Commands

• Run tests: bash test.sh
• Inspect modules: pytest -q tests/test_basic.py
• Lint (optional): flake8 (if configured in the environment)

Repository Rules

• Do not modify the public API in breaking ways without updating tests.
• Each feature is implemented as a small, testable unit under src/gridguard_secure_attested_cross_domain_e.
• The MVP is designed to be extended by other agents and adapters in subsequent sprint phases.

Contributing

• Implement new adapters by adding entries to AdaptersMarketplace and exposing simple interfaces for attestation and policy enforcement.
• Extend Verifiable Optimization with more realistic ZK-proofs and privacy-preserving computations in future iterations.