What problem does it solve?
Help teams convert confusing or incorrect Sanity Agent Context behavior into a concise set of verified Instructions that the agent cannot infer from the schema alone, reducing incorrect GROQ results and improving agent relevance.
Core Features & Use Cases
- Interactive schema dialogue: Guide a conversation to identify which document types matter, misleading field names, and second-order reference chains.
- Verified evidence: Produce every claim with a supporting GROQ query and result, and require user confirmation before drafting instructions.
- Safe deployment paths: Support both write-access workflows (draft context doc) and URL-param testing for environments without write access.
- Use case: Fix an agent that returns irrelevant documents by adding a groqFilter and a short Instructions block that captures data quirks and required filters.
Quick Start
Create concise, verified Custom instructions and a groqFilter for my Sanity Agent Context MCP to resolve schema confusion and improve query accuracy.