AI pentesting
Can You Trust an AI Pentest That Needs Multiple Runs?
Why AI pentesting struggles with consistent vulnerability coverage, why repeated runs don't scale, and how Cascade makes autonomous security testing more reliable.
Yacine Souam
Yacine Souam — Last Publications
Security Research
Jev and the Future of AI Pentesting: Why the Harness Matters More Than Ever
Security teams now use agents for a wide range of work, from finding vulnerabilities in code to investigating and triaging issues. Many of these tasks need complex reasoning and an understanding of how an application works, so the shift makes sense. We're also seeing more security use cases where an LLM may be more than the job needs. Triage is a good example. We ask an LLM whether a finding is a false positive, give it the context it needs, extract its decision as JSON, and use that to trigger
Research
AI vs AI: How Cascade exploited an AI agent in production
We pointed Cascade, Escape's AI pentesting engine, at a production AI agent. The agent had a guardrail built to catch prompt injection. Cascade got past it on the second try. The trick wasn't a smarter payload, just a different pretext, and the agent handed over its entire system prompt: the tools it could call, the rules it followed, and the session details attached to the conversation. That alone gives an attacker what they need to go straight at the agent's tools instead of guessing at them
Agentic Pentesting
How Escape AI Pentesting Exploited SSRF in LiteLLM
At Escape, we routinely test the AI infrastructure that teams deploy inside their cloud environments. LLM gateways, RAG pipelines, model proxies: these are services that make outbound HTTP requests by design, which makes them natural targets for SSRF. When we looked at LiteLLM, we found three confirmed SSRF sinks, a security gate that was specifically built to prevent this exact bug class, and a bypass that means the gate itself is broken. The organization has confirmed to our research team th
