Test Reporting

Parses JUnit XML and test output, extracts docstrings and assertions from your test files, then generates a styled report withWhat/How/Why descriptions for every test, not just failures. Built specifically for multi-agent systems where you need to see the full test harness state, including LLM provider selection, graph specs, and escalation flows. The report is opinionated about structure (component grouping, sticky headers, monospace everything) but saves you from manually digging through pytest output when something breaks in production.