Avoid multisig signer index overflow panic in gRPC compat decode

[Freyskeyd]

Summary

agglayer-interop-grpc-types gRPC v1 compat decode for multisig currently computes entry.index + 1 when sizing the signer vector.

If entry.index == u32::MAX, this overflows and panics in debug/test builds instead of returning a typed validation error.

Agglayer currently needs a local precheck only to avoid that panic before delegating to interop conversion.

Location

• crates/agglayer-interop-grpc-types/src/compat/v1/aggchain_data/v1_to_types.rs

Current Behavior

The code computes:

let required_len = signatures
    .iter()
    .map(|entry| entry.index + 1)
    .max()
    .unwrap_or(0);

This can panic on u32::MAX + 1.

Expected Behavior

• Use checked_add(1) or equivalent safe arithmetic.
• Return a typed Error::invalid_data("Multisig ECDSA signer index overflow") instead of panicking.

Why It Matters

Once this is fixed, Agglayer can rely on interop multisig conversion directly without keeping a local overflow prevalidation step.

Acceptance Criteria

• No panic on entry.index == u32::MAX
• Decode returns a typed invalid-data error
• Add a regression test covering the overflow case