Context
Untrusted inputs (merkle proofs in register, XDR/event decoding, inbound webhook bodies) are prime
fuzz targets and likely under-fuzzed.
Scope
- cargo-fuzz targets for proof verification and any custom decoding.
- JS fuzzing for webhook/payload parsers and input validators.
- Corpus + CI fuzz smoke run; longer scheduled fuzz job.
Acceptance criteria
- Fuzz targets exist for each untrusted parser; CI runs a smoke fuzz; crashes block merge.
Verification
- A seeded panic on malformed proof/payload is caught by the fuzzer.
Context
Untrusted inputs (merkle proofs in
register, XDR/event decoding, inbound webhook bodies) are primefuzz targets and likely under-fuzzed.
Scope
Acceptance criteria
Verification