Add support for Etherscan-style RPC rate-limit error handling / retry logic

[casks-mutters]

When using a public RPC provider (e.g. Infura / Alchemy / free node), it's common to hit rate limits (HTTP 429) or transient network timeouts. README mentions rate-limiting as a possible issue.
Suggested enhancement:

• Add simple retry logic on RPC calls (for block fetch or storage read), e.g. retry up to N times with exponential backoff.
• Make retry behavior configurable (e.g. via --retries, --backoff, --timeout).
• Provide clear error messages when all retries fail.

Benefits:

• Improved resilience for real-world usage.
• Better reliability for large batch runs / CI / scheduled audits.