prakrutimonali

I'm Prakruti, a developer working at the intersection of agentic AI and governance the boring-but-critical layer that stops "let the agent handle it" from turning into a production incident. Right now I'm building agent-governance-starter-kit, an open-source set of policy-as-code templates and a CLI validator for governing AI agents in production because most teams shipping agents today have no repeatable way to check what an agent is allowed to do, who approved it, or what happens when it goes wrong. This blog is where I write about that world: agent governance, observability, policy-as-code patterns, and the practical, often-unglamorous engineering problems that show up once you move agents from a demo to something real. Expect deep dives, build logs, and the occasional "here's what broke and why." Always happy to talk agents, governance tooling, or open source, feel free to reach out.