Case studies

Evidence from real repositories.

Each one links to a public pull request you can read.

These projects are examples of code reviewed with arnx. They are not customers or partners, and none of them endorsed arnx. Every page keeps three things apart: what arnx reported, what was sent upstream, and what the maintainers did on their own. Where a claim rests on our local review records and not on a public page, it says so.

• SQLFluff: valid ClickHouse projections were unparsable

  A generated test showed valid ClickHouse projection SQL was rejected by the parser. Fix merged upstream; arnx's one confirmed finding in the same review was wrong.

• mark: a second Close returned before shutdown finished

  A generated test showed a second concurrent Close returned while the resource was still stopping. Fix merged upstream. Found with Go support that is in test and not released.

• beets: shell completion crashed with the titlecase plugin

  A generated test showed a plugin added the same command-line option twice, which crashed shell completion. Fix open upstream, not merged; most other findings in the review were intended behavior.

• Hummingbot: order statuses paired with the wrong order

  One finding sent upstream as an open pull request, and a separate set of findings that matched fixes the maintainers had already made on their own. arnx caused none of those fixes.

• arnx self-review: remove could delete files outside the repository

  arnx reviewing its own code: a generated test showed arnx remove would delete a file outside the repository. The test still guards that behavior in the public source.

A private repository

From a finding to a checked fix: a Python and Postgres review follows one owner-run review through triage, an agent's fix and the final check. Its source is private, so it shows the workflow, not evidence you can replay.

For how often findings hold up on code we do not know, read the research notes.