skills.rest
Aradotso
AradotsoOFFICIAL
Explore all skills by this author →
Website𝕏@aradotsoGitHub

mcp-code-execution-mode

Proxy MCP tool calls through a sandboxed run_python bridge.

7|2|Updated May 16, 2026
Search Tags:#python#mcp#token efficiency#tool discovery#sandboxing#container runtime#tool proxy
One-click install
npx skills add https://github.com/Aradotso/mcp-skills --skill mcp-code-execution-mode
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: mcp-code-execution-mode
Source: https://github.com/Aradotso/mcp-skills/tree/main/skills/mcp-code-execution-mode
Command: npx skills add https://github.com/Aradotso/mcp-skills --skill mcp-code-execution-mode

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill prevents large MCP tool schemas from ballooning your LLM prompt context by routing work through a single, token-efficient Python execution bridge instead of repeatedly loading many server tool definitions.

Core Features & Use Cases

  • Constant token overhead: keeps prompt size near-constant (~200 tokens) regardless of how many MCP servers you connect.
  • Discovery-first MCP proxying: dynamically queries tool schemas and searches docs only when needed, so you don’t pay the context cost upfront.
  • Rootless, sandboxed execution: runs Python in rootless containers with hardened settings (no network, read-only root filesystem) to reduce risk while enabling practical automation.
  • Persistent sessions: maintains variables and MCP client state across calls within the same session for iterative workflows.

Quick Start

Configure Claude Desktop to use a single MCP server called code-execution, then ask it to run a Python snippet that discovers the needed MCP tools and performs the task (for example, fetch data from a configured server, transform it in Python, and write results back).

Frequently Asked Questions about mcp-code-execution-mode

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I reduce MCP tool schema token bloat in my LLM prompt context?▼

To reduce MCP tool schema token bloat, route agent logic through a single run_python bridge that proxies calls to other MCP servers, keeping prompt size near-constant at roughly 200 tokens regardless of connected servers.

How does a Python execution bridge proxy calls to multiple MCP servers?▼

A Python execution bridge proxies MCP calls by dynamically querying tool schemas and searching documentation only when needed, maintaining variables and MCP client state across calls within the same session for iterative workflows.

Do I need a container runtime like Podman or Docker to run sandboxed Python via MCP?▼

Yes, you need a rootless container runtime like Podman or Docker to run sandboxed Python via MCP, which executes code with hardened settings including no network access and a read-only root filesystem to reduce risk.

What is the best way to dynamically discover MCP tools without loading upfront context costs?▼

The best way to dynamically discover MCP tools without upfront context costs is using a discovery-first proxying approach that queries tool schemas on demand, preventing thousands of schema tokens from loading each turn.

Can I maintain persistent variables and MCP client state across multiple Python calls?▼

Yes, you can maintain persistent variables and MCP client state across multiple Python calls within the same session, enabling iterative workflows such as data analysis and cross-server orchestration without losing state.

What are the limitations of running Python in a rootless sandboxed container?▼

Limitations of running Python in a rootless sandboxed container include no network access and a read-only root filesystem, which restricts operations requiring external connectivity or persistent local file writes outside permitted paths.

Related Skills

Explore More Software Engineering Skills→
kubiyabotkubiyabot
16

docker

Automate Docker container and image management via the native Docker CLI.

Official
Advanced
GitmaxdGitmaxd
7

droid-mode

Access MCP tools with progressive schema disclosure and on-demand hydration.

Community
Advanced
gqf2008gqf2008

native-mcp

Discover MCP tools from external servers and register them as native Hermez Agent tools.

Community
Intermediate
rafaelcallejarafaelcalleja
2

mcp-management

Coordinate MCP servers to discover, analyze and execute their capabilities without polluting the main context.

Community
Advanced