What is the Model Context Protocol (MCP)?

MCP is an open standard from Anthropic that lets AI agents call external tools over a structured JSON-RPC interface. Instead of embedding tool logic inside the model, you run a local MCP server and your AI client connects to it. AgenticStore MCP Tools is an MCP server you can run yourself.

Are AgenticStore MCP Tools really free?

Yes. The open-source toolkit is MIT-licensed and free forever. You can run it locally with no account, no API key, and no usage limits. The commercial AgenticStore platform (cloud-hosted, metered billing) is separate and in private beta.

Do the tools require an internet connection?

Most tools are fully offline — python_lint_checker, repo_scanner, and dependency_audit run entirely on your machine. The one exception is agentic_web_search, which makes an outbound HTTP request to the URL you provide, and dependency_audit, which queries the OSV database to check for CVEs.

Which AI clients are supported?

Any client that supports MCP over stdio. This includes Claude Desktop, Cursor, Windsurf, VS Code (with the MCP extension), and any custom agent built with the MCP SDK.

Do I need Python installed on my machine?

No. If you use uvx (the recommended method), uv manages a fully isolated Python environment automatically. You only need uv itself installed — nothing else.

How do I load only specific tools?

Use the --modules or --tools flag. For example: uvx agentic-store-mcp --modules code loads only AgenticCode tools, and uvx agentic-store-mcp --tools agentic_web_search loads a single tool. See the Quickstart for full examples.

AgenticStore MCP Tools

A free, open-source toolkit that gives your AI agents real capabilities — Python linting, secret scanning, dependency auditing, and web intelligence. Self-hosted over stdio. No account. No API keys for core tools.

Quickstart guide →

Install and run in under 5 minutes

Claude Desktop config →

Connect to Claude Desktop

Cursor config →

Connect to Cursor IDE

GitHub →

Source code + issues

Tool reference

AgenticCode

python_lint_checker

Static analysis for Python — bugs, style, complexity.

repo_scanner

Detect leaked secrets, PII, and .gitignore gaps.

dependency_audit

Audit deps for CVEs and outdated packages.

AgenticData

agentic_web_search

SEO signal extractor — metadata, OG, headings, links.

design principles

Pure Python

No external API calls. No managed services. Every tool runs entirely on your machine.

Modular

Load only the modules you need. Each tool is independent — mix and match freely.

MCP-native

Built for the Model Context Protocol. Works with any MCP-compatible AI client.

Frequently asked questions