Research papers on agentic provenance, execution assurance, trust boundaries and verifiable evidence for autonomous AI systems.
The Agentic Provenance Protocol (APP) research collection examines how independently sourced evidence can connect authorization, delegation, agent execution and observed outcomes. These papers are written for AI infrastructure teams, security engineers, architects and researchers evaluating autonomous systems. Developed by INEVRI, they describe conceptual approaches and architectural research, not a finalized protocol specification or validated production performance. Each publication is available in full below, including its technical examples, evidence limitations and references where provided.
INEVRI Research
Agentic Provenance Protocol (APP)
The missing trust boundary in agentic commerce
An examination of the distinction between authenticated authorization and independent execution provenance across agentic commerce systems.
Read publicationINEVRI Technical Research
Agentic Provenance Protocol (APP)
A reference architecture for cross-system observation, correlation and execution assurance
A proposed architecture for cross-boundary observation, evidence provenance, causal relationships, trust evaluation and independently inspectable evidence receipts.
Read publicationINEVRI Technical Research
Agentic Provenance Protocol (APP)
Evidence states, probabilistic inference and independent assurance in autonomous systems
A source-aware evidence model separating observations, assertions and inference, with uncertainty-aware evaluation and defensible Evidence Receipts.
Read publication