AI security researcher. Red-teaming tool-using LLM agents, grading what they actually do rather than what they say, and evaluating the frameworks that test them
Pinned Loading
-
adversa-guardrail-degradation
adversa-guardrail-degradation PublicAutomated red-teaming framework that measures how LLM safety guardrails degrade under multi-turn adversarial pressure
Python 2
-
action-graded-severity
action-graded-severity PublicAction-graded severity scoring (L0-L6) for tool-using AI agents, computed from red-team execution traces.
Python 2
Something went wrong, please refresh the page to try again.
If the problem persists, check the GitHub status page or contact support.
If the problem persists, check the GitHub status page or contact support.
