agentveil 0.7.10

Python SDK for AgentVeil: agent identity, action gates, signed receipts, and proof packets

AgentVeil

Action control for autonomous agents — check posture, gate risky actions, prove execution.

AgentVeil is the Python SDK for agent action control: posture checks, Runtime Gate decisions, signed receipts, W3C verifiable credentials, plus DID identity, reputation signals, and MCP integrations.

pip install agentveil

Quick Start

Run locally with real cryptography and mocked HTTP. No server is required.

from agentveil import AVPAgent

agent = AVPAgent.create(mock=True, name="demo-agent")
agent.register(display_name="Demo Agent")

rep = agent.get_reputation()
print(rep["score"], rep["interpretation"])

For production setup, see the Customer Integration guide.

What AgentVeil Provides

• Posture checks before risky agent actions reach production.
• Runtime Gate decisions for allow, approval required, or block outcomes.
• Signed receipts and proof packets for audit and offline verification.
• W3C VC v2.0 credentials with eddsa-jcs-2022 Data Integrity proofs.
• DID identity with portable did:key Ed25519 keys.
• Framework integrations for CrewAI, LangGraph, AutoGen, OpenAI, Claude MCP, Gemini, PydanticAI, Paperclip, and AWS Bedrock.

Offline Verification

Fetch a W3C Verifiable Credential:

curl https://agentveil.dev/v1/reputation/{agent_did}/credential?format=w3c

Verify it with any VC library, or with the SDK:

cred = agent.get_reputation_credential(format="w3c")
assert AVPAgent.verify_w3c_credential(cred)

MCP Server

The base install includes the MCP runtime dependency:

pip install agentveil
agentveil-mcp

The compatibility extra agentveil[mcp] still works for legacy setups. MCP setup details are in the MCP README.

Resources

• Full GitHub README and demo
• API reference
• Customer integration guide
• Framework integrations
• Security context
• Examples
• AgentVeil API
• Live Network

License

MIT. See the license.