RTK Security Labs

RTK Security Labs is an independent AI security practice specializing in adversarial testing of large language model (LLM) systems before they reach production. We run adversarial campaigns against deployed and pre-deployment LLM systems, and every engagement produces a signed, independently verifiable evidence artifact.

Campaigns target the prompt-injection and system-prompt-extraction surface (OWASP LLM01) and the sensitive-information-disclosure and data-leakage surface (OWASP LLM02). The artifact records exactly what was tested and what was observed -a clean result and a finding are documented with equal rigor- and any third party can verify it against a published key without relying on us. Results are signed with ECDSA P-256 over RFC 8785 canonicalization with SHA-256.

On the public record: RTK Security Labs' founder filed written comments to the U.S. federal AI procurement rulemaking (GSA-GSAR-2026-0331, August 2026), proposing benchmark construction-quality requirements for federal AI procurement, and spoke at the GSA AI Policy Listening Session at George Washington University Law School (July 2026). Methodology is anchored to the OWASP LLM Top 10 and the NIST AI Risk Management Framework.

Scope: RTK Security Labs focuses on the behavioral-security layer of AI systems -how a model can be manipulated, leaked from, or driven outside its intended behavior. We do not perform data-handling compliance audits, statutory or jurisdictional governance review, or traditional network penetration testing; we partner or refer where those are needed.