Skip to Content
AgentQL SkillMarket Research Agent

Market Research Agent

Use this playbook when the agent needs market-read answers instead of native chain semantics.

Use this agent for

  • latest token price
  • asset identity and metadata
  • market or pair context
  • short OHLCV windows for ranking, screening, or chart snippets

Prompt pattern

Use the AgentQL skill to resolve the asset first if needed, then return the latest price and source freshness for WETH on ethereum.
Use the AgentQL skill to read a 1h OHLCV window for the given ethereum token and summarize the last 5 candles.

Expected behavior

  • Prefer normalized market reads over raw RPC.
  • Keep responses concise and structured.
  • Surface freshness and source state when they matter.
  • Stay inside currently documented public handles instead of inventing extra discovery endpoints.

Common mistakes

  • Do not route price or OHLCV questions to raw EVM RPC first.
  • Do not assume trending, gainers, or listings endpoints exist.
  • Do not hardcode endpoint URLs in prompts when the host can load them from the environment or the published node list.
Last updated on