I publish papers alongside the code, each with a DOI and free to read. This page also lists the standards work I'm part of, independent research that cites the Agent Passport System, press coverage, and my public comments to NIST.
Nine papers, each with a DOI. Full bibliography on ORCID.
The foundational paper. Agent Passport, Human Values Floor and Beneficiary Attribution, the three layers the protocol grew from.
The invariant that keeps delegation chains safe. Authority can narrow as it passes down, never widen.
A product lattice model for how scoped authority narrows across delegation hops, facet by facet.
Governing what a persistent agent may learn and reuse from what it observes, not only what it may do.
Binding hardware attestation to delegation chains. Seven experiments on IBM Quantum hardware.
Why the unit of agent governance is the population and the medium that carries its actions, not the single agent.
Signed commitments over which interpretable model features were active during a decision, tested against Llama 3.1 8B.
Signed receipts show a procedure was followed. They do not show the effect was safe. Five classes of what they miss.
Peer-voted retention kept a worse shared memory than random selection in 4 of 5 replicates. Pre-registered, with kill conditions.
The protocol's interoperability surface for agent identity, delegation and enforcement. I am the author.
An approved lab for testing agent authority across implementations. I proposed it and maintain it.
Reviewer and contributor. The AST09 execution receipts guidance carries a denied-before-dispatch example vector I contributed, merged by the project leader.
Co-lead of the RFC-189 conformance cases.
Independent work that cites or builds against the Agent Passport System. A citation is not an endorsement, and some of these are critiques.
S. Mih and A. Sokolov. Names me in its Contributors section for the content-derived action reference pattern.
Elena Tănase, AIDAXIA. Cites draft-pidlisnyi-aps in related work.
Lifei Liu, Haoran Yu and Danping Zhang. Cites draft-pidlisnyi-aps among recent proposals that combine notification, transparency logs or enforcement records.
Lifei Liu, Haoran Yu and Xiaochong Jiang. Uses APS as a baseline and reports results against it.
Guangjun Liu. Cites draft-pidlisnyi-aps for attenuating authority along finer facets.
Juan Figuera. Receiver-attested receipts for AI agent actions. References the Agent Passport System.
Peter Le Fol du Taillis. Its signed passport envelope follows the pattern of the agent passport proposals, APS among them.
Two drafts reached complete on fields. draft-pidlisnyi-aps is one of them.
Frans Moore. Names delegation governance by APS as one of four composed layers.
Poke-nushi. Lists APS among its closest adjacent work, with a comparison document.
Mayur Agnihotri. Cites the APS Internet-Draft for its reversibility scheme.
References the APS earned-reputation model in section 7.6.
Quotes me on keeping agent authority bounded across systems that do not share a trust model.
Self-certified to ISO/IEC 5230 and ISO/IEC 18974, with the checklists published alongside the source.
The first press coverage of the protocol.
A critique of the passport layer landscape that engages the receipts work directly.
Points to the budget reservation crosswalk in the Agent Governance Vocabulary.
Íñigo Medina added the paper to his research library with an independent summary and critique.
Filed as public comment. Filing is not endorsement.
Comment filed and acknowledged.
Comment on agent identity and delegation primitives.
Concept paper submitted.