- LOC: ~27,300
- Languages: OCaml, Rust, ReScript, Lean4, Idris2, Zig
- Existing ABI proofs:
src/abi/*.idr(template-level) - Existing verification:
docs/proofs/verification/WokeLang.lean— Lean4 proofs exist, 12sorryoccurrences previously eliminated - Dangerous patterns: None remaining (Lean4 file mentions sorry elimination is complete)
WokeLang.leanclaims all 12sorryeliminated — verify this is still true after any subsequent changes- Audit: do the Lean4 proofs cover the full type system or only a subset?
ast.ml,eval.ml,main.ml— the runtime evaluator- If Lean4 proofs cover the type system but not evaluation, there is a gap
- Prove: evaluation semantics match the Lean4 specification
- Compilation to WASM should preserve the properties proven in Lean4
- Prove: WASM codegen produces programs with the same observable behaviour
fuzz/fuzz_lexer.ml,fuzz/fuzz_parser.ml,fuzz/fuzz_targets/fuzz_input.rs- Fuzzing is a good complement but does not replace proofs for the core semantics
- Lean4 (already in use — extend to cover evaluation and compilation)
- Idris2 for ABI layer
MEDIUM — Good existing proof coverage in Lean4. Focus on extending proofs to evaluation and WASM compilation rather than starting from scratch. The sorry-free status needs periodic re-verification.