The Agent Skills Directory

[COMMAND_EXECUTION]: The references/audit-patterns.md file contains specialized shell commands using ripgrep to help users and agents identify RPC usage patterns, loops containing network calls, and potential exposure of provider API keys in environment variables.
[EXTERNAL_DOWNLOADS]: The benchmark utility (found in benchmarks/bench-evm.mjs and benchmarks/bench-solana.mjs) makes legitimate network requests to external blockchain RPC providers, such as Alchemy, Infura, and public nodes, to verify cost and latency benchmarks.
[PROMPT_INJECTION]: The skill instructions define procedures for ingesting and summarizing external blockchain data, which creates an indirect prompt injection surface where the agent processes data from external RPC endpoints; however, the skill provides advice on managing this data through slicing and file-based storage to minimize context risks.

rpc-optimize