The problem
Once a Claude Code session ends, the interactive Q&A moments (where Claude asked questions and I made decisions) are buried in JSONL transcript files. Existing transcript viewers treat everything equally: tool calls, code generation, and decision-making Q&A all get the same weight. The signal gets lost.
What it does
A Node.js script that scans all Claude Code session logs across every project, extracts every AskUserQuestion interaction, and generates a self-contained HTML file. Uses a 3-tier visual hierarchy: Q&A cards front and center, conversation context around them, and tool calls collapsed into pills.
Light/dark mode with system preference detection, expandable conversation entries, JSON export for Q&A pairs, multi-project aggregation, and sticky navigation. Runs standalone or as a Claude Code skill. Everything stays local.