Start here
Read your first pull request score
Open the evidence behind quality, effort, velocity, and AI attribution.
- Audience
- engineering leaders, developers
- Required role
- Any member with Measure access.
- Product access
- Measure
- Navigation path
- Measure > Pull requests > select a pull request
Outcome
Explain what a pull request score says, trace it to evidence, and identify incomplete analysis.
Before you begin
- A repository with at least one processed pull request
Open a scored pull request
Open Measure and select a pull request that shows completed analysis. The detail page connects the summary metrics to commits, files, authorship evidence, and the original GitHub record.
Read the score summary
Quality reflects the rubric-based assessment of correctness and maintainability. Effort estimates how much of the change is meaningful work rather than churn. Velocity combines deterministic change size with the effort multiplier. Read these together, not as independent quotas.
Inspect evidence
Expand the score explanation and review the commits, changed paths, exclusions, and AI attribution confidence. Use the GitHub link to compare source evidence when a result is surprising.
Resolve an unexpected score
Confirm that the pull request is merged, all commits are synchronized, and analysis is complete. Administrators can review scoring settings or request reanalysis after correcting source data. Use feedback from the pull request page when the evidence is complete but the explanation still looks wrong.