Research Papers
Versioned research outputs with explicit publication status, downloadable papers and reproducibility materials.
Vertical AI Governance
Extensible research sequence on semantic sovereignty, executable governance and governed autonomous systems.
“Models may change. Operational knowledge must remain sovereign.”
Governance Conformance Observatory
Empirical audit series measuring framework-native agent primitives against the Vertical AI Governance Stack v1.0 benchmark.
“Observing framework defaults under deterministic proof obligations.”
All Publications
Index of preprints, canonical specifications, draft specifications, instruments, and applied audits.
Failure Semantics, Cancellation, and Side-Effect Idempotency in AI Agent Runners
Empirical evaluation of failure semantics, cancellation recovery, and side-effect idempotency across OpenAI Agents SDK releases v0.20.0, v0.21.1, and v0.22.3 across 366 total trials.
Cross-Version Evaluation of Guardrail Execution Precedence and External Side-Effect Commitment in Agent Runners
Empirical evaluation of guardrail execution precedence and HTTP side-effect dispatch across OpenAI Agents SDK releases v0.20.0, v0.21.1, and v0.22.3 across 2,900 total trials.
Empirical Evaluation of Guardrail Execution Precedence in Asynchronous Agent Runners: An Audit of OpenAI Agents SDK v0.22.3
Empirical evaluation of guardrail execution precedence (happens-before guarantees) in OpenAI Agents SDK v0.22.3 across 1,200 trials under parallel and sequential execution modes.
Vertical AI Governance Stack v1.1 (Draft Specification)
Pre-registered v1.1 draft specification, 22-requirement matrix, dual conformance model (Path A vs Path B), adversarial falsification, and frozen human validation protocol (External Evaluation Pending).
Applied Audit 02: OpenAI Agents SDK
Experimental conformance audit of OpenAI Agents SDK runtime primitives, guardrails, tracing spans, and handoff authorization under Vertical AI Governance Stack v1.0.
Applied Audit 01: OpenAI Swarm
Framework-native governance controls versus externally required controls under the Vertical AI Governance Stack v1.0. Baseline case study in minimal orchestrators.
Vertical AI Governance Stack v1.0 — First Applied Validation
Governance assessment of PrintPriceOS Control Plane exposing execution drift, asynchronous CVA bypasses, and framework candidate refinements.
Vertical AI Governance Stack v1.0
Reference Architecture for Semantic Sovereignty, Invariant Matrix (L1-L6), and Proof Obligations.
Implementation Conformance Checklist v1.0
Operational audit tool for evaluating L1-L6 compliance and the Autonomy Ladder (A0-A4).
Vertical AI Manifesto
Semantic Sovereignty, Executable Governance, and domain-specialized AI infrastructure.
The Oracle Blind Spot
Telemetry verification, test-oracle quality and transport-level re-adjudication.