TKOResearch
Menu

Proof and trust

Review our methods and qualifications.

Explore our review methodology, sample reports, security practices, research, and engineering background before scoping an engagement.

Overview

Materials to support your review.

  • Methodology summary
  • Sample artifacts
  • Security and data handling practices
  • Founder background
  • Program and support disclosures
  • Publication and research links
  • Certification scope and limitation language

What you receive

Review materials

Use these materials to evaluate our approach, review sample deliverables, and identify any additional information your diligence process requires.

Methodology and artifacts

See how we define the systems and permissions under review, document findings, and recommend fixes.

  • Approach page
  • Sample review package
  • Enterprise buyer packet

Founder background

Founder-led review by Kevin O'Connor, with security-engineering and advanced threat research experience.

  • About
  • Media and speaking

Security practices

Security contact details, a PGP key, vulnerability reporting instructions, and NDA and data-handling arrangements before detailed review. Keep sensitive materials out of public forms.

  • Security
  • PGP
  • Security diligence FAQ

Published analysis

Open research and field notes cover AI-agent authority, MCP/tool risk, RAG boundaries, prompt injection, and production-readiness review.

  • Insights
  • OWASP Top 10 for AI
  • Field Notes

01

How we support review work

Methodology

We agree on the decision and systems to review, examine permissions and failure scenarios, and link findings to supporting records and recommended changes.

Background

Kevin O'Connor's background in security engineering, advanced threat research and AI-security assessment.

Programs and support

Program participation and institutional support are distinct from certification of a client system. Ask about the scope of any affiliation relevant to your review.

02

Where findings apply

  • Certification results apply to the defined systems, versions and test date.
  • No guarantee that an AI system is safe in every configuration.
  • A narrow review supports conclusions only about the controls examined.

Next step

Discuss your diligence requirements.

Request diligence materials