What problem does it solve?
This Skill helps developers perform token-efficient, multi-file codebase analysis by delegating work to a dedicated MCP cursor-agent server. It enables broad searches, architecture mapping, and flow tracing across large codebases without manual, file-by-file reading. Use this in contexts where a single-file answer would be expensive or when you need cross-file insights, not just context-limited explanations.
Core Features & Use Cases
- Batch multi-file queries: Submit batched questions to the MCP server and receive consolidated file:line references, code snippets, and purposes.
- Background, asynchronous analysis: Leverages a cursor-agent server to keep your session responsive while long analyses run.
- Cross-file tracing & architecture mapping: Understand how components interact across files and modules, reducing exploration time.
- Use Case: When you need to understand how a feature is implemented across multiple files or when tracing a data flow that spans modules, this Skill provides concrete locations and context.
Quick Start
Start a batched query to locate where a symbol is used across the codebase, e.g.:
start = mcp__cursor_agent__cursor_agent_start({ "query": "Find where X is used. For each location: file:line, code snippet, and purpose." })
Wait for completion with mcp__cursor_agent__cursor_agent_result and then present the findings.
