Function paid_measureTriggerRate

Measure whether a skill's description actually FIRES (recall + precision) by running real prompts against a real model. Spends money.

⚠️ paid_ overstates slightly: an injected evalDriver replaces the billed path, and stubSkillBodies makes the default path much cheaper without making it free. The DEFAULT path bills.

paid_measureTriggerRate(
spec: TriggerRateSpec,
opts?: { evalDriver?: EvalDriver },
): Promise<TriggerRateReport>
Measure a skill/behaviour's real trigger rate across prompts × trials. Defaults to the real claude CLI (claudeEvalDriver); pass { evalDriver } to drive a second harness — e.g. measureTriggerRate(spec, { evalDriver: codexEvalDriver }) from vigiles/codex (the eval-tier analog of runHarnessTest's { adapter }). Requires that harness's binary + auth.
Parameters
- spec: TriggerRateSpec
- opts: { evalDriver?: EvalDriver } = {}
Returns Promise<TriggerRateReport>
- Defined in src/eval.ts:3594

Function paid_measureTriggerRate

Parameters

Returns Promise<TriggerRateReport>

Settings