Response error fields and stable machine-readable error codes returned by Comscore MCP tools.
Every tool response includes these fields from the base result type. On success, error and error_code are null. On failure, error is always populated and error_code carries a stable machine-readable identifier.
| Field | Type | Description |
|---|---|---|
| error | string | null | Human-readable failure message. Set only on failure; null on success. |
| warning | string | null | Recoverable failure message. Includes retry guidance when relevant. May be set alongside a partial result. |
| error_code | string | null | Stable machine-readable code. Use this for programmatic handling — the error text is for humans and may change. null on success. |
| retryable | boolean | null | true when the same call is worth retrying unchanged — indicates a transport or timing failure, not a request error. false or null means the request itself is the problem. |
The following error_code values are stable across releases. All other text in error responses is human-readable and may change.
retryable: true). Retry the same call unchanged.retryable: true). Retry after a short delay.error field includes the API's response for diagnosis. May or may not be retryable depending on context — check the retryable field.error message names a supported alternative when one exists. Not retryable without changing the request. Fix the parameters before retrying.retryable field; if true, the same call may succeed on retry.Social tools that accept batches of entity names include additional fields on the _NamedEntityResult base type for partial resolution failures.
| Field | Type | Description |
|---|---|---|
| unresolved_names | array | null | Names that could not be resolved to a unique entity, with candidates to try. Retry these calls with the specific candidate IDs rather than the original names. |
| resolved_names | array | null | Names that resolved successfully, with their entity IDs. Use these IDs directly on retry rather than re-resolving by name. |
When a batch request partially resolves — some names succeed, some fail — the result includes data for the resolved names and populates unresolved_names for the failures. Inspect resolved_names to extract the IDs for the entities that did resolve, then issue a second call using those IDs for the names that did not.