Agent identity and delegation governance for NeMo agent teams

Is this a new feature, an improvement, or a change to existing functionality?

New Feature

How would you describe the priority of this feature request?

Low (would be nice)

Please provide a clear description of the problem this feature solves

I want NeMo Agent Toolkit to enforce cryptographic identity and scoped delegation between agents in a team, because I need to prevent unauthorized inter-agent actions when Agent A delegates subtasks to Agent B.

Currently, NeMo Agent Toolkit handles orchestration and safety evaluation well, but when agents delegate work to other agents, there is no cryptographic enforcement that Agent B stays within the authority Agent A granted. Agent B operates on trust, not proof. In enterprise multi-agent deployments, this creates an authorization gap: agents can potentially exceed their intended scope without detection at the protocol level.

Describe your ideal solution

A governance middleware layer that sits between agent message dispatch and execution:

1. Each agent gets an Ed25519 identity (keypair + signed metadata)
2. Inter-agent messages carry signed delegation chains proving authorization scope
3. A gateway validates identity + scope before message delivery
4. Every action generates a signed receipt (audit trail)
5. Revoking a parent delegation instantly invalidates all downstream delegations (cascade revocation)
6. Effective permissions are capped by earned trust (reputation-gated authority), not just granted scope
7. Passports issued through official infrastructure are countersigned by the issuer (CA model) — self-signed passports are valid but unverified

The Agent Passport System (Apache 2.0) implements this. Integration with NeMo could work as a policy layer in the agent runtime — intercept tool calls, verify delegation, enforce scope, generate receipts.

Additional context

• SDK: npm install agent-passport-system (v1.29.0, 1852 tests, 62 core + 32 v2 = 94 modules) or pip install agent-passport-system (Python v0.7.0)
• MCP server: 122 tools for any MCP-compatible client
• Issuer CA: passports countersigned by AEOESS issuer key. Published at https://aeoess.com/.well-known/aeoess-issuer.json
• Paper: https://doi.org/10.5281/zenodo.19260073 (Faceted Authority Attenuation — product lattice model)
• GitHub: https://github.com/aeoess/agent-passport-system
• A working group with 4 independent implementations has ratified Entity Verification v1.0, QSP-1 v1.0, and Execution Attestation v1.0

Happy to discuss integration approaches or contribute directly.

Code of Conduct

• I agree to follow this project's Code of Conduct
• I have searched the open feature requests and have found no duplicates for this feature request

No NeMo files, tests, or entry points are named. Start by locating the agent message-dispatch and execution integration points, then compare them with the proposed identity, delegation, validation, revocation, and receipt requirements; done would require an agreed integration design and corresponding validation coverage.