Security researcher and open-source developer building tools around honeypots,
LLM services, and Trusted Execution Environments.
CURRENT PROJECT
A multi-profile honeypot for exposed local LLM services including Ollama, LM Studio, vLLM, Gradio, and ComfyUI. It emulates common LLM API surfaces and records probes and submitted prompts — without running a model or executing attacker input.
Python Docker T-Pot Threat Intelligence
REPRODUCIBILITY ARTIFACT
Artifact for presented research on the complementarity of rule-based and LLM-based bad-partitioning analysis in Trusted Applications.
Program Analysis LLM Security Trusted Execution Environments
Activity
PYTHON · C/C++ · TYPESCRIPT · DOCKER · LINUX