- Home
- Skills
- Yoanbernabeu
- Grepai Skills
- Grepai Mcp Cursor
grepai-mcp-cursor_skill
1
GitHub Stars
1
Bundled Files
2 months ago
Catalog Refreshed
4 months ago
First Indexed
Readme & install
Copy the install command, review bundled files from the catalogue, and read any extended description pulled from the listing source.
Installation
Preview and clipboard use veilstrat where the catalogue uses aiagentskills.
npx veilstrat add skill yoanbernabeu/grepai-skills --skill grepai-mcp-cursor- SKILL.md6.1 KB
Overview
This skill integrates GrepAI with the Cursor IDE using the Model Context Protocol (MCP) to enable semantic code search, call graph analysis, and index-based navigation inside Cursor. It provides configuration examples for per-project and global setups, guidance for workspaces and environment variables, and quick troubleshooting tips. Use this skill to make Cursor's AI invoke GrepAI tools like grepai_search and grepai_trace_* directly from the editor.
How this skill works
You add an MCP server entry pointing to the grepai command (mcp-serve) in .cursor/mcp.json or a global MCP config. Cursor connects to the GrepAI MCP server and exposes tools such as grepai_search, grepai_trace_callers, grepai_trace_callees, grepai_trace_graph, and grepai_index_status. When Cursor's AI receives a code query, it calls these tools to run semantic searches, return call graphs, or check indexing status, letting you navigate code semantically from the IDE.
When to use it
- Enable semantic code search from within Cursor instead of manual CLI calls
- Inspect function callers/callees or visualize call graphs without leaving the editor
- Set up GrepAI integration for a single project or for all projects globally
- Use workspaces to query multiple repositories with a single GrepAI server
- Troubleshoot Cursor-AI not invoking GrepAI tools or empty search results
Best practices
- Put .cursor/mcp.json in the project root for per-project configs and use explicit cwd when needed
- Store environment secrets outside the file; prefer shell profile variables over embedding keys in MCP config
- Keep the GrepAI index fresh with grepai watch --background after large changes or pulls
- Verify grepai is in PATH and test with grepai mcp-serve manually before relying on Cursor
- Use workspaces for multi-repo workflows and set limits on results to improve performance
Example use cases
- Ask Cursor: "Find code that handles user login" — Cursor will call grepai_search to return semantically related files and snippets
- Ask Cursor: "What functions call validateToken?" — Cursor will use grepai_trace_callers to list callers and locations
- Request a call graph: "Show me the call graph for processPayment" — Cursor will invoke grepai_trace_graph and visualize dependencies
- Set up a workspace to search across multiple services with grepai mcp-serve --workspace dev
- Check index health from Cursor using grepai_index_status to confirm search readiness
FAQ
Create .cursor/mcp.json in the project root with a grepai entry pointing to command 'grepai' and args ['mcp-serve'].
How do I make GrepAI available for all projects on my machine?
Add the same mcpServers entry to ~/.cursor/mcp.json (or %APPDATA%\Cursor\mcp.json on Windows) and restart Cursor.
Cursor shows no results — what should I check?
Ensure the project is indexed (run grepai watch), confirm grepai is in PATH, set cwd explicitly in the MCP config if needed, and restart Cursor.