Asking a Tool to Review My Own Diff — Why I Built CommitBrief
One afternoon I hit the limits of self-review, so I wrote a tool that audits my own diff without leaving the terminal: CommitBrief.
Series
From a small CLI that has an LLM review my own diff to a tool the team uses to catch flaky tests: the log of building CommitBrief.
This series has 6 posts.
Part 1
One afternoon I hit the limits of self-review, so I wrote a tool that audits my own diff without leaving the terminal: CommitBrief.
Part 2
Two new commands: summary briefs a change and its history, and remote pr reviews a PR from the terminal, leaving comments before the human does.
Part 3
I added a deterministic flaky-test filter to CommitBrief: it catches hard-coded sleeps and unseeded randomness at commit time, with no model.
Part 4
I measured CommitBrief's default models for the first time: Gemini cleared the threshold, while the OpenAI default raised too many false alarms.
Part 5
In CommitBrief, what counts as the 'same' finding moved from the title to a rule ID; baseline, policy and PR threads were built on that decision.
Part 6
On a partial config, one CommitBrief command silently turned off secret scanning. Silent failures I closed July to October, and the rule: fail closed.