chore(deps): ethers→alloy migration phase 4 (U256/H160 boundary)

[clawdbot-glitch003]

Summary

Phase 4 of the ethers→alloy migration. Migrates the H160/U256 boundary files and the accounts / account_management public APIs to alloy primitives. Codex review surfaced no findings.

Scope-shift from the original plan: the signer_pool / signable_contract middleware stack stays on ethers because the abigen-generated AccountConfig<M> requires M: ethers::Middleware. That swap is now grouped with Phase 5 where the contract bindings are regenerated via sol! and the dependency inverts (alloy contracts take alloy providers). The plan doc is updated to reflect this.

What moved to alloy

• utils/mod.rs, utils/parse_with_hash.rs — keccak256, U256, Address
• core/v1/helpers/api_status.rs — From<ethers FromStrRadixErr> → alloy's ruint::ParseError; dropped unused From<ProviderError> and From<ConversionError> (no callers)
• core/v1/models/curve_type.rs — removed unused TryFrom<ethers::U256> / From<CurveType> for ethers::U256 (alloy's blanket TryFrom on U256 conflicted with the hand-rolled From; no external callers)
• accounts/blockchain_cache.rs — cache-key fns take alloy types; address rendering switched from ethers' lowercase Debug to {:#x} so keys stay deterministic (alloy Debug uses EIP-55 checksum)
• actions/client/op_code_helpers/private_keys.rs — LocalWallet → PrivateKeySigner
• accounts/mod.rs — public API on alloy; bridges to ethers at every contract call site via the new utils::alloy_ethers helpers
• core/account_management.rs — alloy pubsig and internals; wallet construction via PrivateKeySigner. Removes the Phase 3 H160::from_slice(signer.as_slice()) bridge

What stayed on ethers (deferred to Phase 5)

• accounts/signer_pool.rs, accounts/signable_contract.rs — middleware stack stays ethers because contracts require ethers::Middleware
• accounts::Metadata (abigen-generated) — metadata_to_item keeps an ethers::types::U256::zero() comparison until sol! regenerates the struct
• accounts/decode_revert.rs — ethers::contract::ContractError API

New bridge module

src/utils/alloy_ethers.rs — four inline conversion helpers (alloy_u256_to_ethers / ethers_u256_to_alloy / alloy_address_to_ethers / ethers_address_to_alloy) plus a round-trip test. Entire module is removed in Phase 5.

Test plan

• cargo check -p lit-api-server clean (lib + tests)
• cargo clippy --tests -p lit-api-server clean
• 217/217 lib tests pass (215 prior + 2 new bridge round-trip tests; curve_type lost 2 tests with the removed U256 impls)
• codex review --commit HEAD returned no findings
• CI green
• Manual smoke test of one endpoint that exercises a contract write path (e.g. POST /v1/accounts → triggers accounts::new_account and register_wallet_derivation through the new bridge layer)

🤖 Generated with Claude Code

[GTC6244]

Assuming this PR is referencing another PR that will remove all the bridge functions .