It wraps the commands I already use, like pytest, jest, cargo test, and go test, records what failed and what eventually passed, and turns that into a Markdown report. The report includes a summary and the red-to-green window, but it only uses recorded commands, exit codes, and Git changes. It doesn't use AI or guess the root cause.

DebugBrief itself makes no network requests and has no telemetry. It uses the standard library on Python 3.11+, tomli on 3.9 and 3.10, and native Git when available.

pipx install debugbrief

It's currently tested on Linux and macOS with Python 3.9 through 3.14. I'd be interested to know whether this workflow is useful to anyone else.