app: public attestation explorer for independent verification

[joelpeace48-cell]

Problem

The protocol's claim is that anyone can verify an inference attestation. In
practice that currently requires an RPC client and knowledge of the contract
ABI, which means the claim is true only for people who could have verified it
without us.

What to do

A public, no-login route — /explore — where anyone can:

• Look up an attestation by id, transaction hash, or payload hash
• See the full record with its verification material: transaction, ledger,
  contract address, and the model and policy it references
• Follow the supersession chain if the record has been corrected
• Check a payload hash they hold against what is recorded, answering "was
  this output attested" directly
• Link out to a block explorer for independent confirmation

Read through the backend, cache aggressively, and rate-limit — it is
unauthenticated by design.

Acceptance criteria

• Lookup by id, transaction hash, and payload hash
• Full record with verification material displayed
• Supersession chain navigable
• Explorer links for independent confirmation
• Cached and rate-limited

Notes

This is the most persuasive page the site can have. "Paste your hash and check
it yourself" turns a marketing claim into a demonstration, and it is the thing a
sceptical reviewer will actually try.

[prodbycorne]

The core idea is compelling: anyone should be able to independently verify an inference attestation. But requiring an RPC client and knowledge of the contract ABI creates a practical barrier that undermines that claim. A verifier shouldn’t need protocol-specific tooling just to answer a basic question: “Is this attestation valid?”

I’d be motivated to work on this because making verification accessible would significantly improve the protocol’s transparency and credibility. A simple, user-facing verification path would let developers, researchers, auditors, and other users validate attestations independently, without having to understand the underlying contract interface first.

My experience with developer tooling and technical workflows has also shown me that verification is only truly useful when it is easy enough for people to actually perform. This feels like an important usability and trust layer for the protocol, not just a convenience feature.