pdd-to-app-journeys-eval

To validate a CommCare journey document for app deployability, grade it against a 7-dimension LLM-as-Judge rubric. This process checks persona specificity, archetype alignment, edge-case recoverability, and field-ready requirements to ensure the app meets field expert standards.

LLM-as-Judge grading for product design validation is an automated evaluation method that scores journey sets against a rigorous rubric. It assesses coverage completeness, happy-path voice, and pass-criteria measurability to catch gaps in validation that manual review often misses before app build initiation.

To check if your PDD journey sets cover edge cases and input validation, apply a deployability hard gate evaluation. This automatically fails journey sets lacking critical field requirements like GPS accuracy gating, case write-back, and data quality enforcement, even if they are fully faithful to a thin PDD.

Yes, you can use automated journey grading for ACE Connect opportunities. The evaluation validates product design documents by applying a 7-dimension rubric to journey sets generated from idea-to-PDD artifacts, ensuring the final CommCare product meets real-world data collection standards.

A limitation of using an LLM rubric for CommCare deployability checks is that it applies a hard gate for out-of-chain requirements. Journey sets will automatically fail if they lack critical field requirements like input validation or GPS accuracy gating, regardless of their overall fidelity to the PDD.

No, the pdd-to-app-journeys-eval skill does not require any external dependencies. It operates independently to grade product design documents and journey sets against its built-in 7-dimension LLM-as-Judge rubric to deliver a deployability verdict.