ad-tdg

Outcome-based prompting is a workflow that applies a known canonical technique to generate a ground-truth pair, a Test Dependency Map, and three candidate approaches to land a verified implementation.

To generate a Test Dependency Map, you define a ground-truth pair and document the dependencies required for verification, ensuring a repeatable workflow for your selected implementation strategy.

Use the ground-truth pair workflow when the canonical technique is known but implementation is uncertain, and route to exploratory spikes only when the underlying technique itself remains uncertain.

To select a single evaluation criterion for three implementation approaches, draft the trade-offs of each approach and commit to one criterion to verify the final result with an auditable trail.

A documented verification loop enforces generating a ground-truth pair and Test Dependency Map, drafting three approaches, and applying a single-criterion selection to commit a verified implementation.

No, if the technique itself is uncertain, this workflow routes you to an exploratory spike instead, as it is specifically designed for technique-known but implementation-uncertain scenarios.