Looking for arXiv cs.AI endorsement — independent researcher, LLM agent governance paper

Hi everyone,

I am an independent researcher seeking an arXiv cs.AI endorsement for my first preprint.

Paper: “Policy-Based Agentic Systems (PBAS): Where the Model Proposes and the System Governs”

The paper introduces PBAS — an architectural framework that formally separates planning from execution in LLM-based agent systems. The central contribution is a deterministic Policy Engine that evaluates every proposed plan against an organizational policy matrix before any execution begins. Permissions are enforced at the API layer, not via prompt instructions — making the system structurally resistant to prompt injection.

The reference implementation (DAF) is fully open-source:

github.com

GitHub - tonyos-lab/daf: Deterministic Agentic Framework — Policy-Based...

Deterministic Agentic Framework — Policy-Based Agentic Systems (PBAS)

v0.1.0 — Apache 2.0 — 561 tests passing including 53 adversarial security tests.

Happy to share the PDF on request.

Endorsement code: 8TAABS

Thank you for considering — any help is greatly appreciated.

Tony Ochinang tony.ochinang@tonyos-lab.org