fix(lido): fix panic in eth_call — reqwest::blocking inside async runtime (v0.2.1)

[GeoGu360]

Problem

eth_call used reqwest::blocking::Client, which internally creates its own tokio runtime. Calling it from within a #[tokio::main] async context causes an immediate panic:

thread 'main' panicked: Cannot drop a runtime in a context where blocking is not allowed.
This happens when a runtime is dropped from within an asynchronous context.

Affected commands (all call eth_call from async context):

Command	Call sites
balance	2 (balanceOf, sharesOf)
stake	1 (isStakingPaused pre-flight check)
get-withdrawals	2 (getWithdrawalRequests, getWithdrawalStatus)
claim-withdrawal	2 (getLastCheckpointIndex, findCheckpointHints)

Fix

• Convert eth_call to async fn in onchainos.rs
• Replace reqwest::blocking::Client with async reqwest::Client
• Add .await to all 7 call sites across balance.rs, stake.rs, get_withdrawals.rs, claim_withdrawal.rs
• Bump version to 0.2.1 in Cargo.toml, SKILL.md, plugin.yaml, .claude-plugin/plugin.json

Test plan

• lido get-apy — APR query (unaffected, returns 2.39%)
• lido balance — on-chain stETH balance query, no panic
• lido get-withdrawals — lists withdrawal requests from chain, no panic
• lido request-withdrawal --amount-eth 0.00001 --confirm — 2-step tx (approve + requestWithdrawals), both confirmed on-chain
• lido claim-withdrawal --ids <ID> — reaches chain correctly (returns expected RequestNotFoundOrNotFinalized for PENDING requests, confirming RPC call succeeds)

🤖 Generated with Claude Code

[GeoGu360]

Superseded by #93 — rebased on fork main to avoid .github/ diff false positive.