Explain retained quality measurements and their example counts
/v1/evaluations/runs/{runId}/scorecardRead-only server-derived scorecard for one immutable Evaluation run. Requires current parent Evaluation run-view access. Sums retained integer counts, never averages percentages, invokes targets, regrades, edits labels, or grants approval. Active or legacy runs without intact terminal evidence remain unmeasured. Corrupt evidence suppresses numeric claims. Percentages describe the selected examples only; no population confidence interval is inferred without a sampling protocol. Optional metricId returns a bounded page of contributions, without outputs or reference labels. No-store.
How to call this endpoint
Every ACP API request uses bearer authentication. The examples here show the actual request path, auth header, and body shape that the platform expects.
Path, query, and header parameters
These parameters control which ACP object the endpoint acts on and how the request is processed.
| Name | Location | Type | Required | Description |
|---|---|---|---|---|
| runId | path | string | Yes | — |
| Name | Location | Type | Required | Description |
|---|---|---|---|---|
| metricId | query | string | No | Exact metric ID from this scorecard; omit to return no example bodies. |
| limit | query | integer | No | — |
| offset | query | integer | No | — |
Body schema
This endpoint does not require a request body.
What the API returns
Each response code below includes the documented payload shape for the ACP API.
| Field | Type | Required | Description |
|---|---|---|---|
| scorecard | object | Yes | — |
| scorecard.schemaVersion | computer_agents_quality_scorecard_v1 | Yes | — |
| scorecard.runId | string | Yes | — |
| scorecard.evaluationId | string | Yes | — |
| scorecard.evaluationVersionId | string | Yes | — |
| scorecard.targetType | string | Yes | — |
| scorecard.targetId | string | Yes | — |
| scorecard.targetVersionId | string | Yes | — |
| scorecard.purpose | string | Yes | — |
| scorecard.executionStatus | string | Yes | — |
| scorecard.measurementState | measured | partial | unmeasured | invalid | Yes | — |
| scorecard.decisionState | not_assessed | Yes | This read model never authorizes activation. |
| scorecard.trustLevel | verified_worker | self_reported | unknown | Yes | — |
| scorecard.evidenceFingerprint | string | Yes | — |
| scorecard.datasetFingerprint | string | Yes | — |
| scorecard.caseSelectionFingerprint | string | Yes | — |
| scorecard.evaluatorFingerprint | string | Yes | — |
| scorecard.populationCount | integer | Yes | — |
| scorecard.metrics | object[] | Yes | — |
| scorecard.metrics[].id | string | Yes | Unique identifier. |
| scorecard.metrics[].label | string | Yes | — |
| scorecard.metrics[].description | string | Yes | Human-readable description. |
| scorecard.metrics[].numeratorLabel | string | Yes | — |
| scorecard.metrics[].denominatorLabel | string | Yes | — |
| scorecard.metrics[].unit | percent | Yes | — |
| scorecard.metrics[].aggregation | micro | Yes | — |
| scorecard.metrics[].direction | higher_is_better | lower_is_better | informational | Yes | — |
| scorecard.metrics[].state | measured | partial | unmeasured | invalid | Yes | — |
| scorecard.metrics[].value | number | Yes | A percentage |
| scorecard.metrics[].numerator | integer | Yes | — |
| scorecard.metrics[].denominator | integer | Yes | — |
| scorecard.metrics[].populationCount | integer | Yes | — |
| scorecard.metrics[].includedCount | integer | Yes | — |
| scorecard.metrics[].excludedCounts | object | Yes | — |
| scorecard.metrics[].excludedCounts.unmeasured | integer | No | — |
| scorecard.metrics[].excludedCounts.missing_reference | integer | No | — |
| scorecard.metrics[].excludedCounts.source_changed | integer | No | — |
| scorecard.metrics[].excludedCounts.not_applicable | integer | No | — |
| scorecard.metrics[].excludedCounts.error | integer | No | — |
| scorecard.metrics[].excludedCounts.skipped | integer | No | — |
| scorecard.metrics[].excludedCounts.missing_output | integer | No | — |
| scorecard.metrics[].excludedCounts.invalid | integer | No | — |
| scorecard.metrics[].uncertainty | object | Yes | — |
| scorecard.metrics[].uncertainty.kind | not_estimated | Yes | — |
| scorecard.metrics[].uncertainty.reason | string | Yes | — |
| scorecard.warnings | string[] | Yes | — |
| scorecard.fingerprint | string | Yes | Stable across example pagination; binds this projection and its evidence identity. |
| scorecard.examples | object | Yes | — |
| scorecard.examples.metricId | string | Yes | — |
| scorecard.examples.data | object[] | Yes | — |
| scorecard.examples.data[].caseId | string | Yes | — |
| scorecard.examples.data[].name | string | Yes | Human-readable name. |
| scorecard.examples.data[].state | measured | unmeasured | missing_reference | source_changed | not_applicable | error | skipped | missing_output | invalid | Yes | — |
| scorecard.examples.data[].numerator | integer | Yes | — |
| scorecard.examples.data[].denominator | integer | Yes | — |
| scorecard.examples.hasMore | boolean | Yes | — |
| scorecard.examples.nextOffset | integer | Yes | — |