I work on agent security and evaluation: tool-abuse detection, prompt injection through retrieved content, and trust-boundary failures across multi-agent handoffs. Trace-level rather than per-request.
Building Fluiq getfluiq.com. Also maintain polygate, an open-source unified LLM client, on PyPI and npm.
If you're running agents in production I'd genuinely like to hear what breaks. fluiqai@gmail.com