Tauri / Rust Full-Stack Developer Specialist

Most AI coding help behaves like autocomplete. This skill makes your agent behave like the engineer accountable for the whole desktop product.

When it's active, the agent inspects the real project before it edits, verifies the installed Tauri and Rust versions instead of assuming APIs still exist, and treats the WebView frontend as untrusted ‚AI validating and authorising every command in Rust, applying least-privilege capabilities and a restrictive Content Security Policy. It reasons across every affected layer: requirement, UI/UX, frontend state, the typed IPC contract, the Rust command adapter, domain logic, persistence and migrations, integrations, errors and diagnostics, security, tests, packaging and documentation. Crucially, it is honest about evidence. It distinguishes verified, reviewed, not-verified, failed and blocked results, and it will not claim "all tests pass", "secure" or "production-ready" without proof. It covers the parts that break in the real world‚ AI packaged-vs-dev differences, code signing, macOS notarisation, the updater, database migrations and rollback, and platform differences across Windows, macOS and Linux. The result is fewer plausible-but-wrong changes, tighter security at the IPC boundary, and a repeatable path from "code written" to "actually shipped".