mastra-ai/mastra · error · MastraError

MASTR_SCORER_FAILED_TO_RUN_MISSING_GENERATE_SCORE

MASTR_SCORER_FAILED_TO_RUN_MISSING_GENERATE_SCORE

Error message

Cannot execute pipeline without generateScore() step

What it means

MastraScorer builds a pipeline of steps and only permits execute/run after a generateScore step has been added, since the scorer cannot produce results without it. run() performs a runtime check on hasGenerateScore and throws MASTR_SCORER_FAILED_TO_RUN_MISSING_GENERATE_SCORE when the pipeline lacks that step.

Source

Thrown at packages/core/src/evals/base.ts:957

    requestContext: RequestContext | undefined,
    keys: string[] | undefined,
  ): Record<string, unknown> | undefined {
    if (!requestContext || !keys || keys.length === 0) {
      return undefined;
    }

    const safe = requestContext.serializeForSpan();
    const selected = keys.includes('*') ? safe : selectFields(safe, keys);

    return Object.keys(selected).length > 0 ? selected : undefined;
  }

  async run(input: ScorerRun<TInput, TRunOutput>): Promise<ScorerRunResult<TAccumulatedResults, TInput, TRunOutput>> {
    const { _internal, ...scorerInput } = input;

    // Runtime check: execute only allowed after generateScore
    if (!this.hasGenerateScore) {
      throw new MastraError({
        id: 'MASTR_SCORER_FAILED_TO_RUN_MISSING_GENERATE_SCORE',
        domain: ErrorDomain.SCORER,
        category: ErrorCategory.USER,
        text: `Cannot execute pipeline without generateScore() step`,
        details: {
          scorerId: this.config.id ?? this.config.name,
          steps: this.steps.map(s => s.name).join(', '),
        },
      });
    }

    // Apply prepareRun transformation before span creation to reduce data
    // flowing into both the observability span and the scorer pipeline.
    const prepared = this.config.prepareRun ? await this.config.prepareRun(scorerInput) : scorerInput;

    let runId = prepared.runId;
    if (!runId) {
      runId = randomUUID();

View on GitHub (pinned to 75dd419e61)

Solutions

Add a .generateScore(...) step to the scorer's pipeline before running it.
If steps are chained conditionally, make generateScore mandatory/unconditional in the chain.
Check that you are calling run/execute on the fully built scorer, not a partial builder result that skipped the scoring step.

Example fix

// before
const scorer = new Scorer({ id: 's1' })
  .preprocess(preprocessStep);
await scorer.run(input); // throws
// after
const scorer = new Scorer({ id: 's1' })
  .preprocess(preprocessStep)
  .generateScore(scoreStep);
await scorer.run(input);

Defensive patterns

Strategy: validation

Validate before calling

// check the scorer pipeline includes a generateScore step before running
if (!scorer.hasGenerateScore) {
  throw new Error('Scorer pipeline must include generateScore() before run()');
}
await scorer.run(input);

Try / catch

try {
  const result = await scorer.run(input);
} catch (e) {
  if (e instanceof MastraError && e.id === 'MASTR_SCORER_FAILED_TO_RUN_MISSING_GENERATE_SCORE') {
    throw new Error(`Scorer '${e.details?.scorerId}' was built without generateScore()`);
  }
  throw e;
}

Prevention

Always terminate scorer builder chains with .generateScore(...).
Make generateScore unconditional when steps are added programmatically.
If type-safe builder variants exist (scorers without generateScore lacking run), use them so the mistake is a compile error.

When it happens

Trigger: Calling scorer.run(...) (or execute) on a scorer whose builder chain never invoked .generateScore(...) — e.g. only preprocess/other steps were added — then attempting to execute the scorer pipeline.

Common situations: Composing a scorer from steps and forgetting the scoring step; conditionally chaining steps with the generateScore branch omitted; refactoring a scorer and accidentally dropping the generateScore() call.

Related errors

AI-assisted analysis of mastra-ai/mastra@75dd419e61 (2026-08-30). Data as JSON: /api/errors/e010ad2f60748d87. Report an issue: GitHub.