domains / families / agent-tooling / score-eval

score-eval

authentic

Use when scoring an evaluation diff against a supplied rubric and baseline fixture: retain evidence for detected/fixed findings and report unscorable or unexecuted checks.

When to use

• You have an evaluation diff, rubric, and baseline fixture and need to assess detected and fixed problems.

When not to use

• This skill does not provide the upstream evaluation fixture or rubric.
• Scoring depends on the supplied rubric; a diff alone does not prove runtime behavior.
• Result writes and test execution must remain within the user's authorized evaluation scope.

Skill page

Source