← all posts
LaunchMarch 8, 2026·5 min read

Introducing AgenticStore MCP Tools — Free, Open-Source Superpowers for Your AI Agents

We're open-sourcing our first batch of MCP tools. Python linting, secret scanning, dependency auditing, and SEO-grade web search — all self-hosted, no account required.

Today we're publishing AgenticStore MCP Tools v1 — a free, open-source toolkit that gives your AI agents real superpowers. No cloud account. No API keys for core tools. Just install and run.

We built this because we kept running into the same problem: AI agents are incredibly capable, but they're often flying blind. They can write code, but they can't check if it's correct. They can suggest dependency updates, but they don't know which packages have active CVEs. They can fetch a URL, but they can't parse its SEO signals.

The Model Context Protocol (MCP) changes this. It gives AI agents a standard interface to call real tools — file systems, APIs, databases, and now our toolkit. But most MCP tools require accounts, API keys, or send data to external services. We wanted something different: a set of tools that are genuinely useful, completely local, and free forever.

What's in v1

We're shipping four tools across two modules:

AgenticCode
python_lint_checker

Static analysis for Python files — syntax errors, unused imports, complexity. Zero dependencies.

repo_scanner

Scans your repository for leaked secrets, API tokens, PII, and .gitignore gaps. Checks for 40+ credential patterns.

dependency_audit

Audits requirements.txt, package.json, go.mod, pom.xml, and more against the OSV vulnerability database.

AgenticData
agentic_web_search

Fetches any URL and returns a full SEO signal report: title, meta description, canonical URL, robots directives, Open Graph tags, heading structure, word count, and link breakdown.

Why open source

We're building a commercial platform — AgenticStore — that lets developers subscribe to a curated catalog of high-quality tools delivered over a single MCP endpoint. But we don't think the foundation should be proprietary. The tools in this open-source toolkit are pure Python, run locally, and will always be free.

We want AI developers to trust their tooling. You can read every line of code, fork it, modify it, and run it on air-gapped machines. That kind of transparency matters when your AI agent is scanning your codebase for secrets or auditing your dependencies.

How to get started

If you have uv installed, you can be up and running in one command:

bash
uvx agentic-store-mcp

Then add it to your Claude Desktop, Cursor, or Windsurf config and restart. Full instructions are in the quickstart guide.

What's next

v1 is the foundation. We have a clear roadmap: more AgenticCode tools (TypeScript/JavaScript support, test coverage analysis), AgenticData expansions (structured data extraction, site crawling), and AgenticOps tools for infrastructure and deployment tasks.

If you want to contribute — or if there's a tool you wish your AI agent had — open an issue on GitHub. Every tool we accept has to be pure Python, local-only, and genuinely useful.

And if you want early access to the full AgenticStore platform — where we handle the infrastructure, scaling, and billing so you can focus on building — join the waitlist.

Frequently asked questions