Latest Articles

Insights and guides on AI security and LLM application development

Scanning the OpenAI Cookbook with Kereva-Scanner

OpenAI maintains a community contributed 'Cookbook' of patterns for using their API. We scanned it for code issues.

March 20, 2025Andrew Marble

Deterministic scans of AI model implementations

Beyond evals: A look at scanning tools for AI model security and performance.

March 15, 2025Andrew Marble

Evals are not all you need

Evaluations are useful for ranking models but flawed as a guarantee of software performance.

March 12, 2025Andrew Marble