Run both audits below in order. Report findings in a structured table. Iterate with the lemmafit-proofs skill (Step 4) until only minor findings remain.

Dafny code patterns and reference for lemmafit apps. Use when writing or editing .dfy files, defining state machines (Model, Action, Inv, Init, Step), or when Dafny verification fails and you need to fix errors. Covers the AppCore module pattern and common mistakes.

Generate human-readable guarantees from proven Dafny code and verify them with claimcheck. Use after verification succeeds and SPEC.yaml is in sync with Dafny. Produces guarantees.json, claimcheck-mapping.json, runs claimcheck-multi, and generates guarantees.md report.

Audit workflow for lemmafit apps after React. Runs after writing React code to catch logic that slipped through. Performs audit and labels findings by severity and iterates until only minor findings remain.

Proof-maximizing workflow for writing Dafny lemmas, postconditions, and invariants. Load this skill before writing any lemmas, ensures clauses, or when deciding between axioms and lemmas. Covers postcondition-first development, scaffolding, biconditional completeness, and input validation patterns.

Lemmafit development workflow for managing SPEC.yaml and Dafny specifications. Use when the user describes a new feature, asks to add or update features or specs, or when spec changes need to be addressed. Handles SPEC.yaml entries, Dafny code sync, and spec queue processing. Follow these instructions whenever your workflow touches the spec.yaml file