AI Prompt Fuzzer
Summary: PortSwigger Research lists a Burp extension called AI Prompt Fuzzer, authored by Mohamed Idris, that automates prompt fuzzing against AI APIs using customizable payloads to find prompt injection vulnerabilities. Version 3.0.0, last updated 04 September 2025, adds an AI vs AI mode in which Burp AI generates payloads, analyzes and evaluates responses, and decides the next action, consuming more Burp AI credits.
What you should know: The tool flags responses containing specified validation strings, and its own documentation notes that LLM responses are non-deterministic, recommending payloads that ask the model to echo specific phrases when it accepts a prompt. That is the detection problem a security leader inherits: a prompt injection finding depends on the model repeating a marker back, not on a deterministic pass or fail.
Read the original at PortSwigger Research ↗
- Topics
- AI Security
- Related
- SECURITY AFFAIRS AI-CYBERSECURITY NEWSLETTER ROUND 8 Cyber Risks · Nov 15, 2026Pentagon Renames CDAO ‘Super Intelligence Office’ Cyber Risks · Nov 15, 2026186 Fake AI Domains Redirect Users to Click Fix Lures, Browser Malware, and Search Hijackers Cyber Risks · Nov 12, 2026