Press and articles
- The Agent Times: "Someone Just Shipped the Credential Layer We’ve All Been Missing". The first press coverage of the protocol.
- OPAQUE (PR Newswire, via Yahoo Finance), 2026-09-28: “OPAQUE's Open Source Approach to Verifiable AI Nears Half a Million Downloads as Community Code Contributions Grow More Than Tenfold.” The release quotes Tymofii Pidlisnyi, author of the Agent Passport System, on keeping agent authority bounded across systems that do not share a trust model, and on TRACE making runtime evidence verifiable outside the system that produced it.
- Michael Chen (dev.to), 2026-05-23: "The Agent Spend Governance Gap". On agent spend limits as a governance problem, pointing to the budget reservation crosswalk incubating in the Agent Governance Vocabulary as the upstream canonical verb set.
Contributions to other projects
- XERJ Corpus Hub, 2026-10-28: the open reference index for coding agents lists Tymofii Pidlisnyi as a contributor, with the Cycles receipt-to-envelope implementation from the APS SDK selected as a worked reference, under Apache 2.0 with attribution.
- OWASP Agentic Skills Top 10, v1.0 released 2026-08-17: project page and whitepaper (PDF). The AST09 execution receipts guidance describes the bilateral receipt pattern, an admission receipt before execution and an outcome receipt after, keyed by attempt_id, with a denied-before-dispatch example vector contributed by APS (PR #45, merged by the project leader).
Research references and criticism
- Notarized Agents (Juan Figuera, arXiv:2606.04193): receiver-attested confidential receipts for AI agent actions; references the Agent Passport System.
- Composable Trust Infrastructure for the Agent Internet (Frans Moore, 10.5281/zenodo.19263547): a federated architecture composing four layers, naming delegation governance by the Agent Passport System as one of them.
- Verifiable Agent Trust Envelope (Poke-nushi, 10.5281/zenodo.19839769): lists the Agent Passport System among its closest adjacent work, with a dedicated comparison document.
- The Governability Gap in Agentic AI (Mayur Agnihotri, 10.5281/zenodo.21376745): why authority gating cannot be proven across composed multi-agent flows; cites the APS Internet-Draft for its reversibility scheme.
- PDR in Production (University of British Columbia): references the APS earned-reputation model.
- Íñigo Medina (Instituto Tramontana), 2026-04-19: catalogued the working paper "Governance in the Medium" in his research library, with an independent summary and critique.
- Agent Authority Manifest (AAM) (Gaurav Arora, Institute for Technology Stewardship), 2026-11-10: a public consultation draft, 10.5281/zenodo.24104750, open for comments until 2027-01-15. It names the Agent Passport System Internet-Draft as the closest related work and says AAM should reuse its restriction mechanics.
- AgentAction, itself a project in this field, 2026-11-13: "The AI Agent Governance Landscape" lists the Agent Passport System among independent, solo-maintained projects and calls it conceptually the most complete design in that tier, while noting that its academic framing is self-published, not peer-reviewed.
- Armorer Labs: "Every agent passport layer is grading its own exam", a critique of the passport-layer landscape that engages the receipts work directly.
None of these is an endorsement of APS.
OpenChain self-certification
- OpenChain Project, 2026-07-17: "Agent Passport System Joins the OpenChain Community of Conformance". APS self-certified to ISO/IEC 5230:2020 for open source license compliance and ISO/IEC 18974:2023 for security assurance; the completed checklists are published with the source.
Boilerplate
AI agents act for people and organizations, but the systems around them often cannot prove the basics: who authorized the agent, what it was allowed to do, and what happened. Agent Passport System is an Apache 2.0 open-source protocol for verifiable agent authority. It binds an agent to the principal it represents, constrains delegated authority by scope, spend, time and other conditions, and produces signed evidence of permit and deny decisions at the enforcement boundary. The result is a cryptographic trail connecting authority, enforcement and action. The enforcement gateway is open source as well. EU rules for high-risk AI systems, as amended by Regulation (EU) 2026/1744, increasingly require logging and traceability, although that regulation deferred when those obligations apply.
Facts
| License | Code Apache 2.0 |
| Spec | draft-pidlisnyi-aps (IETF Internet-Draft) |
| Editor | Tymofii Pidlisnyi |
| Research | Papers and citations |
Contact
Interviews, technical clarifications, or embargoed material: [email protected].