Authors & Contributors

The people building and improving AIRQ — an independent map of AI agent security risk for enterprise buyers, security teams, and vendors.

1 About the Project

AIRQ (AI Risk Quadrant) is a vendor-neutral research program that scores production AI agents on Attack Surface, Blast Radius, and Defense Controls, then places them on a shared risk map.

The framework is defined in the Methodology and applied to the current agent landscape in the AI Risk Quadrant Report. The methodology evolves independently from each report edition as the attack surface changes.

AIRQ is produced by AI security researchers and practitioners. Contributions are welcome, credited where appropriate, and reviewed against the published scoring rubrics.

2 Contributors

The authors, reviewers, and supporters behind the AIRQ framework and report.

  • Eugene Neelou AI Security Office, Adversa AI
  • Serge Malenkovich Adversa AI
  • Alex Polyakov Coalition for Secure AI, Adversa AI
  • Tiffany Saade Cisco AI Defense
  • Paolo Di Prodi CrowdStrike
  • Ken Huang Cloud Security Alliance
  • Bill Stout Coalition for Secure AI
  • Sarah Novotny Coalition for Secure AI
  • Apostol Vassilev NIST
  • Om Narayan OWASP
  • Emmanuel Guilherme OWASP

3 How to Contribute

We accept evidence-backed corrections, methodology proposals, and implementation partnerships. Every submission should cite primary sources the scoring team can verify.

  • Submit agent corrections — flag a scoring error, report a missing CVE, or update a defense status with evidence.
  • Improve the AIRQ methodology — propose changes to attack surface weights, defense criteria, or the AIRQ Score formula.
  • Partner on AIRQ implementation — integrate AIRQ scores into procurement, compliance, or security workflows.
  • Free-form proposal — new agent nominations, class boundary suggestions, research collaborations, or anything else worth discussing.

Ready to send a correction, proposal, or partnership note? Use the form below — we read every message and follow up when a submission needs more detail or affects a published score.