AGA MCP Server

Cryptographic runtime governance for AI agents. 20 tools. Sealed policy artifacts, continuous measurement, tamper-evident proof. Ed25519 + SHA-256.

AGA - Attested Governance Artifacts

Cryptographic runtime governance for AI agents and autonomous systems.

npm PyPI License: MIT Tests

# Try it now
pip install aga-governance
python -m aga demo
python -m aga verify demo-bundle.json

What This Does

Every tool call an AI agent makes passes through the AGA gateway. Each call is evaluated against policy, and the decision (PERMITTED or DENIED) is recorded as a signed, hash-linked governance receipt. Receipts are collected into evidence bundles that any third party can verify offline using standard cryptography.

Record. Prove. Verify.

Use with Claude Desktop

Add to your Claude Desktop MCP config (claude_desktop_config.json):

{
  "mcpServers": {
    "aga": {
      "command": "npx",
      "args": ["-y", "@attested-intelligence/aga-mcp-server"]
    }
  }
}

Claude can then seal artifacts, measure integrity, generate evidence bundles, and verify compliance through natural language.

MCP Tools (20)

CategoryTools
Identityget_server_info, get_portal_state
Lifecycleinit_chain, attest_subject, revoke_artifact
Enforcementmeasure_integrity, measure_behavior, verify_chain
Evidencecreate_checkpoint, generate_evidence_bundle, verify_bundle_offline
Privacyrequest_claim, list_claims
Delegationdelegate_to_subagent
Auditget_receipts, get_chain_events

Quick Start

Verify an evidence bundle (3 commands)

pip install aga-governance
curl -s https://aga-mcp-gateway.attestedintelligence.workers.dev/bundle -o evidence-bundle.json
python -m aga verify evidence-bundle.json

Or verify in your browser

Go to attestedintelligence.com/verify and click "Run Verification." Zero installs required.

How It Works

AI Agent                  AGA Gateway                    Verifier
   |                          |                              |
   |-- tools/call ----------->|                              |
   |                    [Evaluate Policy]                    |
   |                    [Sign Receipt]                       |
   |                    [Chain to Previous]                  |
   |<-- PERMITTED/DENIED -----|                              |
   |                          |                              |
   |                    [Export Bundle]                       |
   |                          |--------- evidence.json ----->|
   |                          |                  [Verify Signatures]
   |                          |                  [Verify Chain]
   |                          |                  [Verify Merkle Tree]
   |                          |                  [PASS / FAIL]

MCP Governance Proxy

Run AGA as a transparent proxy between any MCP client and any MCP server. Every tool call gets evaluated against policy and produces a signed receipt.

# Start the proxy with an upstream MCP server
npx tsx src/proxy/index.ts start --upstream "npx -y @modelcontextprotocol/server-filesystem /tmp/test" --profile standard

# Export the evidence bundle
npx tsx src/proxy/index.ts export --output evidence.json

# Verify
npx tsx src/proxy/index.ts verify evidence.json

The proxy intercepts tools/call requests, evaluates them against a sealed policy artifact, and generates signed receipts. Permitted calls are forwarded to the downstream server. Denied calls return an MCP error. Every decision is hash-linked into a tamper-evident chain.

Three built-in policy profiles:

  • permissive - log everything, block nothing (default)
  • standard - rate limits + blocks destructive operations
  • restrictive - explicit tool allowlist, all unknown tools denied

Verification (5 steps)

  1. Algorithm Check - Bundle declares Ed25519-SHA256-JCS, fail closed on anything else
  2. Receipt Signatures - Ed25519 over RFC 8785 canonical JSON (signature field excluded)
  3. Chain Integrity - Each receipt's previous_receipt_hash = SHA-256 of the preceding receipt
  4. Merkle Proofs - Walk siblings/directions to root, compare against bundle root
  5. Bundle Consistency - Proof count = receipt count, leaf hashes match receipt hashes

Cryptographic Primitives

PrimitivePurpose
Ed25519Receipt signatures
SHA-256Hash chaining, Merkle trees, leaf computation
RFC 8785 (JCS)Canonical JSON for deterministic signing
Merkle TreesBinding all receipts to a single verifiable root

Live Gateway

The demo gateway is deployed on Cloudflare Workers:

# Check status
curl https://aga-mcp-gateway.attestedintelligence.workers.dev/health

# Export evidence bundle
curl https://aga-mcp-gateway.attestedintelligence.workers.dev/bundle -o evidence-bundle.json

Python SDK

pip install aga-governance
from aga import AgentSession

with AgentSession(gateway_id="my-gateway") as session:
    session.record_tool_call(
        tool_name="search_web",
        decision="PERMITTED",
        reason="tool in allowlist",
        request_id="req-1",
    )
    bundle = session.export_bundle()
    result = session.verify()
    assert result["overall_valid"]

Test Suite

355+ automated tests across TypeScript and Python:

  • TypeScript MCP Server: 218 tests (vitest)
  • Python SDK: 137 tests (pytest)
  • Cross-language test vectors: 37 vectors across 9 categories
npm test                              # TypeScript tests

For the Python SDK, install aga-governance from PyPI: https://pypi.org/project/aga-governance/

Project Structure

src/                   # Core protocol: artifacts, receipts, chain, Merkle, crypto, portal state machine
  core/                # Governance primitives (artifact, receipt, chain, portal, bundle)
  crypto/              # Ed25519, SHA-256, BLAKE2b, Merkle, JCS canonicalization
  proxy/               # MCP governance proxy (transparent interception + policy enforcement)
  tools/               # MCP tool handlers (20 tools)
  middleware/          # Zero-trust governance enforcement wrapper
independent-verifier/  # Standalone verifier with zero AGA imports
scenarios/             # Deployment scenarios (SCADA, drone, AI agent)
tests/                 # TypeScript test suite (218 tests)

Links

Security

See SECURITY.md for vulnerability reporting.

Contributing

See CONTRIBUTING.md for development setup and guidelines.

License

MIT


Attested Intelligence Holdings LLC

Related Servers

NotebookLM Web Importer

Import web pages and YouTube videos to NotebookLM with one click. Trusted by 200,000+ users.

Install Chrome Extension