Thanks for the PR. Current triage suggests: **ask for CI fix**. Reviewability score: **0/100**. Before the next maintainer pass, please: - Get CI passing or explain why the failing check is unrelated. - Add a short validation or test plan to the PR body. - Resolve merge conflicts before requesting another review pass. - Consider splitting the PR or explaining why the current scope needs to stay together. _Generated as a draft with Maintainer Radar. Please edit before posting._
Local PR triage for maintainers
Maintainer Radar turns pull request metadata into a deterministic queue brief: what is ready for review, what is blocked by CI, and why each score changed. One command can recommend the next maintainer workflow, create a read-only GitHub Action report, or build a time-boxed review plan that appears in the run summary, with no hosted service or model token.
maintainer-radar recommend https://github.com/owner/repo/pulls maintainer-radar init-repo --profile balanced
Try a public repo
Browser preview. The CLI gives deeper hydrated scans.1 PR has unresolved maintainer blocker.
Workflow: blocker-sweepClear maintainer blockers, merge conflicts, failing CI, or merge gates before assigning review time.
Next 60 minutes: handle 2 PRs in about 17 minutes; 1 quick unblock.
30 minute review plan
17m active - 13m openDraft follow-ups
1 editable ask| PR | Action | Next step | Score | Risk impact | Signals |
|---|---|---|---|---|---|
| #42 Fix parser cache race | review now | Review now while the PR appears small, active, and low risk. | 100 | CI passed (-8 risk) | CI passed, review required, mergeable, review requested, test plan present, tests changed |
| #43 Add universal plugin system | ask for CI fix | Ask the author to get failing checks green before deeper review. | 0 | very large diff (+30 risk); CI failing (+30 risk); merge conflicts (+20 risk) | merge conflicts, changes requested, stale 22 days, no test plan found |
The CLI stays boring on purpose
It reads GitHub data through gh or offline JSON, then emits
Markdown, JSON, CSV, or standalone HTML. The scoring is transparent
enough to inspect and small enough to adjust for a repository's workflow.