Agents call APIs, spend money, and talk to other agents on your behalf. But who said they could? Agent Passport gives every AI agent a cryptographically signed token proving what it's allowed to do — scoped permissions, spend limits, time windows, and delegation chains that can only narrow, never escalate. One npm install. Three lines of code. Works with MCP, LangChain, and any agent framework.
Hey PH! I built Agent Passport because every AI agent framework lets agents DO things, but none of them ask "should this agent be allowed to?" One rogue tool call can send emails, delete files, or spend money without your consent. Agent Passport fixes that — real Ed25519 crypto, not just vibes. It's open source, free, and takes 3 lines to add to any agent. Would love your feedback on what integrations to build next.
Report
No reviews yetBe the first to leave a review for Agent Passport