I build AI-driven security tooling and show the work: live demos, running code, and the judgement to know where the machine stops and the engineer starts.
A five-agent newsroom that researches, drafts, reviews and publishes my posts, gated by four taps from my phone.
Scoring Semgrep findings by probability-of-real, cheaply enough to run inside a pull-request check.
My self-hosted Hermes-Agent: persistent memory, MCP tools, and the runtime under Scribe.
Coding agents commit under developer credentials, so the log names someone who never saw the change. Decide whose name goes on the risk register before you decide what an agent may merge.
Review everything, mentor the prompts, never let it merge alone.
Generated code enters your estate like a dependency — treat it with the same provenance discipline.
The 90-day rollout plan, the one-page policy, and the twelve questions to put to every AI-tooling vendor — assembled into a working document.
Get the guide →