Purpose
Agent Passport MCP is the MCP server of the Agent Passport System, positioned as the enforcement and accountability layer for AI agents, supporting bring your own identity. It provides identity, delegation, enforcement, and signed receipts as MCP tools, used to issue passports, verify narrowing-only delegation chains, enforce gateway policies, and issue signed records.
Main capabilities
The default configuration is full, providing all 152 tools, covering categories such as identity, delegation, enforcement, commerce, reputation, governance, coordination, and data. By setting APS_PROFILE=essential, a streamlined configuration of 25 tools can be used, covering the basic primitives required by most integrations (identity, delegation, enforcement, commerce, reputation). Available configurations include: full (default), essential, identity, governance, coordination, commerce, data, gateway, comms, minimal.
Tools are divided by layer. Some of the categories and counts listed in the materials include: identity (Layer 1) 5 tools, coordination (Layer 6) 11 tools, delegation (Layer 1) 4 tools, Agora (Layer 4) 6 tools, values/policy (Layers 2 & 5) 4 tools, commerce (Layer 8) 3 tools, communication (Agent-to-Agent) 4 tools, agent context (enforcement middleware) 3 tools, principal identity 6 tools, reputation-gated permissions 5 tools, agent gateway 6 tools, intent network (agent intermediary matching) 6 tools.
The protocol surface includes Wave 1 accountability primitives: Ed25519 ActionReceipt, AuthorityBoundaryReceipt, CustodyReceipt, ContestabilityReceipt, APSBundle, as well as strict RFC 8785 JCS for interoperable receipts, with byte matching across implementations.
Usage
The fastest way is a remote connection (no installation required): connect via SSE to mcp.aeoess.com/sse, zero dependencies, and restart the AI client. You can also automatically configure Claude Desktop and Cursor, then restart the AI client. Or add it to the MCP configuration file; the remote SSE method is the same.
Command-line run: npx agent-passport-system-mcp.
Version correspondence: agent-passport-system-mcp 6.0.0 corresponds to SDK 6.0.0; 5.0.2 corresponds to SDK 5.x.
Documentation for AI agents: visit agent-passport.org/llms.txt for machine-readable documentation, or llms-full.txt for the complete technical reference. MCP discovery: .well-known/mcp.json.
Applicable scenarios
Suitable for any MCP client, including Claude Desktop, Claude Code, Cursor, Windsurf, etc. Suitable for integration scenarios that need to establish identity for AI agents, delegation chain verification, enforcement policies, and signed receipt accountability.
Related ecosystem and limitations
Provides multilingual SDKs: npm SDK agent-passport-system (v7.2.1), Python SDK agent-passport-system (v4.2.1), Rust SDK agent-passport-system (v0.3.0, library crate is agent_passport), Go SDK agent-passport-go (v0.7.0, go get github.com/aeoess/agent-passport-go@v0.7.0).
The related project agent-passport-access-shim is an adapter that issues a signed AccessReceipt for each governed MCP tools/call; receipts can be verified using the SDK or in the browser at agent-passport.org/verify.html.
The materials also mention: three contribution PRs were merged by Microsoft maintainers into the Microsoft Agent Governance Toolkit (#274, #598, #1328); a submission was made to NIST NCCoE regarding AI Agent Identity andconfigure Claude Desktop and Cursor, then restart the AI client. Or add it to the MCP configuration file; the remote SSE method is the same.
Command-line run: npx agent-passport-system-mcp.
Version correspondence: agent-passport-system-mcp 6.0.0 corresponds to SDK 6.0.0; 5.0.2 corresponds to SDK 5.x.
Documentation for AI agents: visit agent-passport.org/llms.txt for machine-readable documentation, or llms-full.txt for the complete technical reference. MCP discovery: .well-known/mcp.json.
Applicable scenarios
Suitable for any MCP client, including Claude Desktop, Claude Code, Cursor, Windsurf, etc. Suitable for integration scenarios that need to establish identity for AI agents, delegation chain verification, enforcement policies, and signed receipt accountability.
Related ecosystem and limitations
Provides multilingual SDKs: npm SDK agent-passport-system (v7.2.1), Python SDK agent-passport-system (v4.2.1), Rust SDK agent-passport-system (v0.3.0, library crate is agent_passport), Go SDK agent-passport-go (v0.7.0, go get github.com/aeoess/agent-passport-go@v0.7.0).
The related project agent-passport-access-shim is an adapter that issues a signed AccessReceipt for each governed MCP tools/call; receipts can be verified using the SDK or in the browser at agent-passport.org/verify.html.
The materials also mention: three contribution PRs were merged by Microsoft maintainers into the Microsoft Agent Governance Toolkit (#274, #598, #1328); a submission was made to NIST NCCoE regarding AI Agent Identity and Public comments on the Authorization standard; collaboration with the authors of the IETF DAAP draft on the delegation specification; and multiple papers and an IETF Internet-Draft (draft-pidlisnyi-aps). Documentation is at aeoess.com/llms-full.txt, security information is in SECURITY.md and the GitHub security advisory, and Agora is at aeoess.com/agora.html.
Regarding limitations: the materials do not provide further details on fees, pricing, the scope of platform support, performance results, or specific usage steps, and none are currently provided.
