What’s in a scorecard
Each scorecard evaluates the candidate across four dimensions, derived directly from their comments and actions during the review loop.- Code quality reasoning — How well the candidate identified real issues versus superficial ones. Strong reviewers explain the root cause of a problem and suggest a concrete path forward; weaker reviewers flag style preferences or leave vague observations without substance.
- Risk detection — Whether the candidate caught security vulnerabilities, correctness bugs, and data hazards embedded in the pull request. This dimension reflects the candidate’s ability to recognize code that could cause harm in production, not just code that could be cleaner.
- Revision judgment — How the candidate responded to the AI-generated revision of the pull request. Did they re-read the diff carefully? Did they catch any regressions the revision introduced? Did they update or withdraw comments that were already addressed? This dimension is especially differentiating for senior roles.
- Hiring recommendation — A summary signal calibrated to the difficulty level and specialization settings you configured for the assessment. The recommendation synthesizes performance across the other three dimensions into a single actionable output for your hiring team.
Accessing scorecards
Navigate to the relevant assessment in your Merge dashboard, then select the candidate’s name from the Invites tab. The scorecard opens in a dedicated view that places the candidate’s comments alongside the code, so you can see exactly what they flagged and how they framed it. You can share the scorecard link with other authorized members of your team. Only users within your organization who have been granted access can view scorecard details.Using scorecards in your hiring process
Merge scores are one input among others. Use them alongside interviews, references, and additional signals when evaluating a candidate. No single assessment result should be the sole basis for a hiring decision.
Interpreting Results
Dig deeper into each scorecard dimension and learn how to calibrate expectations by difficulty level.
How the Review Loop Works
Understand the candidate experience — what they see, what the AI revision does, and how scoring is derived.

.png?fit=max&auto=format&n=cbLTSPUrrw39LGR1&q=85&s=dcd381a1cca3b3374f46750c9f794ed8)