Results for “suspicious-flows”

Analyzes prompt injection, retrieval poisoning, memory contamination, planner drift, and tool-boundary abuse in agentic systems, mapping trust boundaries and proving exploit chains.

Detect and exploit DOM-based XSS vulnerabilities by auditing JavaScript for tainted data flow from controllable sources to dangerous sinks, with payloads and bypass techniques for client-side testing.