Upgrade priority:
- High if you use multiplexed mode (--hl-node-compliant-multiplexed) with JWT authentication (--rpc.jwtsecret), or need latest-block read-precompile simulation (#146). Both flags are non-default config.
- Low otherwise.
#146 introduces debug/trace support on the latest block by forwarding read-precompile calls to an upstream RPC when --forward-read-precompiles is specified. This feature is disabled by default.
Requests to official public RPC endpoints (https://rpc.hyperliquid.xyz/evm) are limited to 5 requests per second; custom RPC endpoints are unlimited by default. Configure the endpoint with --read-precompile-rpc-url [hl-node EVM RPC URL] (for example, http://localhost:3001/evm) and optionally set a limit with --read-precompile-rpc-rate-limit [requests-per-second].
Since this is a new feature, please let me know on Discord or Telegram if you encounter any issues. README.md has also been updated.
Also fixes a security issue where multiplexed HTTP/WS RPC could bypass configured JWT authentication (aa29740). The Engine API was unaffected.
What's Changed
- feat(rpc): resolve read precompiles at the chain head through an hl-node by @sprites0 in #146 (Resolves #26)
- fix(rpc): preserve JWT auth and server limits across multiplexed restart by @sprites in aa29740
Full Changelog: nb-20260818...nb-20260904