AI-Powered SDLC Agent Platform
A multi-agent platform that turns a prompt into generated code, tests, design, and an audit trail. This is the append-only debugging journal — honest about what broke, and specific about why.
Newest answer to oldest question
The Cache Trap: When a System Fails Identically Every Time
Two runs of the same prompt produced byte-identical test files — both containing a test that asserted a plus sign would lowercase to False. The system wasn't broken randomly. It was broken reproducibly, which is worse. Plus the architectural decision that removes a whole risk class.
Gated 2026.05.01The Failure Mode Wasn't the LLM — It Was the Contracts Between Agents
Codegen wrote source and the tester independently guessed what was in it. The reviewer judged code it couldn't see. A fallback string became a project requirement and four stages built on top of it. Every implicit assumption between agents was a place a bug could hide.
Gated 2026.05.01Most "AI Flakiness" Was Infrastructure Flakiness
I spent days convinced the model was the problem. Random hangs, empty error messages, runs dying mid-pipeline. It was a file watcher, a timeout, and an event loop policy. When part of your system is non-deterministic, every other bug gets to hide behind it.
Gated 2026.05.01Building an AI-Powered SDLC Agent Platform — The Honest Build Log
Where it started. The architecture, the agent roster, the war stories. What it takes to build an agent platform for the software development lifecycle — and what I'd do differently.
Gated