Grade CommCare Connect app UX screenshots against PDD ground truth.
npx skills add https://github.com/dimagi-internal/ace --skill app-ux-eval
Please help me install this Agent Skill. Skill: app-ux-eval Source: https://github.com/dimagi-internal/ace/tree/main/skills/app-ux-eval Command: npx skills add https://github.com/dimagi-internal/ace --skill app-ux-eval
Manually evaluating the user experience of field worker (FLW) apps is time-consuming, subjective, and often disconnected from original project design requirements. This skill eliminates that gap by providing a structured, rubric-based LLM judgment over captured app screenshots to objectively assess if the built app delivers a good experience for its intended users, with all judgments tied directly to concrete project design document (PDD) ground truth.
Invoke the app-ux-eval skill as part of the /ace:qa-deep workflow to automatically grade the field worker experience of your built app against project design document requirements using the latest captured test journey screenshots.
High-intent search queries and answers about installing and using this skill.
You can grade field worker app UX by evaluating captured test journey screenshots against project design document ground truth, scoring six weighted UX dimensions to produce pass or fail verdicts for app launch readiness.
You can automate QA grading for CommCare Connect builds by running a rubric-based evaluation over captured test journey screenshots to objectively assess if the app delivers a good field worker experience.
Mobile app screenshot analysis scores six core UX dimensions: clarity, flow predictability, error recovery, time budget adherence, journey completion, and capture robustness, applying weighted scores and hard deduction rules for critical failures.
You gate app launch readiness by reviewing per-journey and overall phase pass or fail verdicts generated from the UX evaluation, which explicitly flag critical failures like dead-end errors, incomplete journeys, or time budget overruns.
Unclear field labels and dead-end errors trigger hard deduction rules during the UX evaluation, severely penalizing the rubric score and likely causing the test journey to receive a fail verdict for launch readiness.
dimagi-internalGrade Nova-built CommCare Deliver apps against Product Design Documents for conformance and deployability.
lucifertrjAudit mobile app repositories across iOS, Android, Flutter, and React Native for release readiness.
dimagi-internalGenerate Maestro test recipes from PDD journeys for Nova-built CommCare and Connect apps.
kunalrawat425Audit live developer experience by testing onboarding, docs, and CLI feedback.
