Show HN: AgentSign – Open-source zero trust engine for AI agents

Posted by AskCarX 6 hours ago

Hi HN. This week Meta acquired Moltbook (agent social network), OpenAI acquired Promptfoo (agent testing), and Mandiant's founder raised $190M for Armadin. Agent infrastructure is clearly where things are heading.

We built AgentSign -- a zero trust engine for AI agents. The problem: agents are operating without any identity infrastructure. Moltbook went viral for fake posts because there was zero verification on who or what was posting.

AgentSign gives every agent a cryptographic identity certificate, signs every action into an execution chain, and runs runtime code attestation before anything executes. There's also an MCP Trust Layer for agent-to-MCP server verification, and a Stripe-powered Trust Gate for agent payments.

5 subsystems: identity certs, execution chain verification, runtime code attestation, output tamper detection, and cryptographic trust scoring.

Free and open source. Built in London.

SDK: https://github.com/razashariff/agentsign-sdk

Happy to answer questions.

Comments

Comment by dhblumenfeld1 3 hours ago

V interesting. Does the passport have history of what the agent has done in the past and if so, how are you able to verify that work? Specifically thinking about agents that may work across different systems.

Might be worth checking this out: https://www.8004.org/ it's more crypto specific but trying to tackle the problem of verifying an agent cross platforms.

Comment by socialinteldev 4 hours ago

the identity cert + execution chain combination makes sense -- the real risk with agents isnt just which agent is this but whether you can verify the agent hasnt been injected with malicious instructions between operations

the mcp trust layer is the interesting piece: are you signing each mcp tool call in the chain, or just the agent at initialization? the attack surface shifts a lot depending on when attestation happens relative to the tool call

also curious about the stripe trust gate vs x402 for payment verification -- stripe adds human account overhead; x402 lets agents pay autonomously with on-chain settlement and no human intermediary to impersonate. seems like a better fit for zero-trust agent interactions