This repository contains skill files that help bridge the gap between Open Data Kit (ODK) and the power of AI Agents.
Key Features:
AI-Driven Form Development: Streamline the creation and debugging of complex XLSForms using LLMs.
Server Integration: Frameworks for agents to check submissions and projects on ODK Central.
Data Analysis: Use with Claude Code and OpenCode to perform rapid data synthesis and reporting.
Explore the repo here:
This is great, thanks for sharing! 
I'm Jiayu from Swiss TPH working with @aurdipas and @Thalie. We've been building something that could pair nicely with this - a Claude Code skill for validating existing XLSForms. Where your repo focuses on authoring forms with an AI agent end-to-end, ours focuses narrowly on the QA step: taking an .xlsx a researcher has already drafted and returning plain-English feedback (exact cell, what's wrong, how to fix it) so non-technical field teams can debug without a data manager in the loop.
Under the hood it also wraps pyxform and adds ~10 extra checks for silent-failure patterns we've hit in deployments (broken ${var} scope in repeats, choice-list mismatches, settings-sheet gaps, etc.).
This could work well as a complementary "validator" step after your skill generates a form, before it's pushed to ODK Central.
Please see here: https://github.com/SwissTPH/XLSForm-Validation-Skills
Happy to chat if there's interest in aligning conventions (e.g. your -88/-89/-90 special values) so the two skills play well together 
