Worked on improving documentation reliability and AI-assisted testing quality for the SpacetimeDB Rust SDK repository. Addressed a critical issue where incorrect use of the 'schedule' keyword in table definitions led to hallucinations in large language model-driven tests. By updating the documentation to use the correct 'scheduled' keyword, reduced AI-generated misinformation and improved the determinism of automated test pipelines. This change enhanced the trustworthiness of developer-facing documentation and streamlined onboarding for new contributors. The work involved AI prompt engineering, documentation tooling, and integration with LLM test workflows, utilizing Markdown and Rust to align documentation with the expectations of AI tooling.