Linking Results To Runs

Evaluation results may include draftRunId and publishedRunId.

Those ids reference underlying skill-run records, but the current public docs do not promise /api/runs/:id as a public inspection route.

What to use publicly

• the evaluation result object itself
• diagnosticSummary
• the stable agent runtime if you need a public persisted run-detail API

See also

• Inspect Your First Run
• Public Boundary