fix(avm-transpiler): check u16 operand narrowing

[Kuhai9801]

Summary

Fixes #24115.

The AVM transpiler now checks Brillig memory addresses before encoding them as U16 operands instead of silently wrapping oversized values. The C ABI entrypoints convert transpilation panics into TranspileResult errors, so invalid inputs do not unwind across FFI.

The debug-log message-size immediate is also range-checked before U16 encoding.

Checklist

• I've provided a paragraph or two giving a summary of the change in the description, including relevant motivation and context.
• I've enabled auto-merge if the PR is ready to merge.
• I have reviewed my diff in github, line by line and removed unexpected formatting changes, testing logs, or commented-out code.
• Every change is related to the PR description.
• I have linked this pull request to relevant issues.

Testing

• git diff --check
• Targeted source sweep for remaining to_u32() as u16 and message_size.0 as u16 narrowing

No local compile or test run; verification should run in CI.