A Claude Code skill that records the choices you make while building software — each decision with its options, tradeoffs, and lifecycle phase — and renders them as a searchable, browsable HTML viewer so you can revisit why things were done a certain way.
Project decisions live in people's heads, scattered chat logs, or hidden behind AI. Months later nobody remembers why a choice was made, what else was considered, or what it traded off — so teams re-litigate settled calls and repeat old mistakes.
I built Decision Explorer, a Claude Code skill that captures each decision the moment it's made — the options weighed, their tradeoffs, what was chosen and why — as one small JSON file. It then generates an interactive viewer you open in any browser: browse by recency, lifecycle phase, or release; search; and open any decision to see the full options-vs-tradeoffs comparison and how it changed over time. No build step, no dependencies beyond Python.
Founders almost never log their decisions — with or without AI. The reasoning lives in someone's head until it's forgotten. The challenge wasn't storing decisions; it was getting them captured at all, since no one stops mid-build to document.
A working tool with a live demo, deployed continuously from GitHub. Decisions are logged in plain conversation and instantly explorable — turning scattered reasoning into a durable, browsable record the whole team can revisit.