06 · Services
AI application security assessment
Applications that embed language models: prompt injection, tool and agent abuse, data exfiltration, and the trust boundaries the model creates. We assess AI applications; we are not an AI company.
- OWASP Top 10 for LLM Applications
- MITRE ATLAS
What is tested
Copy pending. Source: 02-services-catalog.md, section for AI application security.
How
Copy pending.
What you receive
Copy pending. Report, readout, and one retest within 90 days are standard (D-008).
Included
Copy pending.
Out of scope by default
Copy pending.