feat(audit): Kick off public security review with formal verification framework

[0xsecaas]

This PR starts a public, developer-led security audit of the threshold-signatures codebase.

Key contributions in this initial phase:

• Added Kani formal proof for transmute_state unsafe cast.
• Introduced audit/README.md and supporting Markdown files to track audit methodology, findings, and risks.
• Documented critical path invariants and potential failure modes for transmute_state.
• Laid out plans for fuzzing, CI checks, and additional formal proofs.

⚠️ Note: This is just the beginning. The audit structure, methodology, and documentation will evolve as we analyze more code, and findings may change significantly.

💡 Feedback and contributions are highly encouraged, especially for identifying attack surfaces and critical paths.

[netrome]

Thank you for creating this PR @0xsecaas. We're currently prioritizing other lines of work and this would require some verification before we can merge it. I.e. we will likely not have the capacity to properly review this anytime soon, so please be aware that this may be on hold for a while.

[0xsecaas]

@netrome Thanks for clarifying, I understand the prioritization.

For context, this PR, besides the audit/ markdown files, addresses a security concern in an unsafe block.

1. I added a unit test to guard against unintended memory layout mismatch, and
2. formally proved the soundness of that unsafe code (a proof that it produces results consistent with the safe u64::from_le_bytes alternative;

This provides additional assurance without changing functionality.