Fix concurrency issue with Ethena claims

[naddison36]

Summary

Fixes a race condition in the automated Ethena withdrawal claim action.

The action previously fetched totalRequests and individual unstaker states using separate latest RPC calls. A concurrent request or claim could therefore produce a mixed snapshot, causing the action to infer an invalid FIFO position and fail with errors such as:

Missing Ethena unstaker 28 for pending request 238

Changes

• Pin all Ethena queue reads to a single block
• Use the same block for adapter state, unstaker mappings, cooldown data, and timestamp checks
• Include the snapshot block number in queue consistency errors
• Update the queue consistency test