Safe In-App MCP Tooling

Community

Safe, bounded MCP tool access in-app.

AuthorJDL440
Version1.0.0
Installs0

System Documentation

What problem does it solve?

This tool provides a safe, app-owned MCP tooling surface for in-app agents by exposing only approved MCP tools and enforcing strict runtime policies to prevent unsafe access.

Core Features & Use Cases

  • App-owned bounded loop: Manage tool access via a registry-based allowlist sourced from mcp/tool-registry.mjs.
  • Prompt contract and runtime guardrails: Clearly declare allowed tools to the agent, require a structured tool request, validate arguments, and cap tool calls to prevent abuse.
  • Production-ready safety: Ensures compatibility with LM Studio and local providers, with default read-only exposure to minimize risk and provide strong fallback behavior for unsafe prompts.

Quick Start

Configure and run the in-app MCP tooling workflow to enable a safe, bounded tool loop for local agents.

Dependency Matrix

Required Modules

None required

Components

Standard package

💻 Claude Code Installation

Recommended: Let Claude install automatically. Simply copy and paste the text below to Claude Code.

Please help me install this Skill:
Name: Safe In-App MCP Tooling
Download link: https://github.com/JDL440/nfl-eval/archive/main.zip#safe-in-app-mcp-tooling

Please download this .zip file, extract it, and install it in the .claude/skills/ directory.
View Source Repository

Agent Skills Search Helper

Install a tiny helper to your Agent, search and equip skill from 471,000+ vetted skills library on demand.