proposal(near): add resiliency path for delayed fast-transfer fee finalization

[omarespejel]

Proposal: NEAR fast-transfer fee claim resiliency path

This is a proposal issue to align on behavior before implementation PRs.

Context

In claim_fee_callback, fee claim for fast-transfer flows currently depends on prior finalization state of the origin leg.

File context:

• near/omni-bridge/src/lib.rs (claim_fee_callback flow)

Proposal goal

Add an operationally safe recovery path so relayer funds do not remain blocked indefinitely when finalization is delayed for non-malicious reasons (prover outage, long chain disruption, etc.).

Candidate design options

1. Timeout-based escape hatch:

• after a configured waiting period, allow governance-controlled resolution path

2. Explicit dispute + resolution path:

• relayer opens claim after timeout, governance can approve/reject with event trail

Guardrails

• keep anti-abuse protections that prevent arbitrary fee claims
• require full event/audit trail on non-happy-path resolution

Test expectations

• normal path unchanged
• timeout path allowed only after threshold
• unauthorized/manual claim attempts rejected

Maintainer input requested

1. Preferred model: timeout-only vs timeout+dispute?
2. Desired ownership: DAO-only or dedicated role?
3. Any SLA targets to encode in config defaults?

Tracked under epic: #556