Security audits
Audits of self-hosted MCP server deployments for real vulnerabilities, not theoretical checklists. Every finding is documented, and responsible disclosure is the standard. Available as the ECHOS methodology matures.
engagements
Work with Botzr Research
Botzr Research takes on a small number of engagements per quarter, focused on the problems the research touches directly.
Benchmark evaluation
If you're building or deploying an AI memory system, Botzr Research can run it through WasTrue, an in-progress benchmark of how memory holds up when facts change. One caveat stated up front: Botzr Research also builds dreamd, a memory layer WasTrue measures, so the report names that conflict of interest and hands you a harness you can rerun yourself.
Technical advisory
Short-term advisory for teams building on MCP, agent memory, or sandboxed execution. The work stays at the system level. Not available for general AI product consulting.
Reach out: pagreen@dreamd.dev