Code Execution Mode MCP Server

Code Execution Mode MCP server enabling AI agents to execute code in multiple programming languages — running Python, JavaScript, and other language snippets, returning execution results and errors, and integrating code execution capabilities into agent-driven programming assistance and code validation workflows.

Evaluated Mar 06, 2026 (0d ago) vcurrent

Homepage ↗ Repo ↗ Developer Tools code-execution mcp-server sandbox python javascript runtime

⚙ Agent Friendliness

/ 100

Can an agent use this?

🔒 Security

/ 100

Is it safe for agents?

⚡ Reliability

/ 100

Does it work consistently?

Score Breakdown

⚙ Agent Friendliness

MCP Quality

Documentation

Error Messages

Auth Simplicity

Rate Limits

🔒 Security

TLS Enforcement

Auth Strength

Scope Granularity

Dep. Hygiene

Secret Handling

No sandboxing — code runs with host permissions. Only run trusted code. No auth or scope controls. Community MCP.

⚡ Reliability

Uptime/SLA

Version Stability

Breaking Changes

Error Recovery

Best When

An agent needs to execute code snippets and validate results locally — for coding assistance where the agent writes and tests code iteratively.

Avoid When

You need production-grade sandboxing, container isolation, or execution of untrusted code — use mcp-sandbox or similar Docker-based solutions.

Use Cases

• Executing Python code snippets from coding assistant agents
• Running JavaScript/Node.js code from web development agents
• Validating code correctness before deployment from testing agents
• Performing data analysis with live code execution from analysis agents
• Testing algorithm implementations from algorithm design agents
• Running code examples from documentation and tutorial agents

Not For

• Production code execution (local execution without full sandboxing)
• Long-running processes or servers (designed for snippet execution)
• Untrusted code from external sources (minimal sandboxing)

Interface

REST API

GraphQL

gRPC

MCP Server

Yes ↗

SDK

Webhooks

Authentication

Methods: none

OAuth: No Scopes: No

No authentication — local MCP server executing code on the host system. Code runs with process owner permissions. No sandboxing beyond process isolation.

Pricing

Model: free

Free tier: Yes

Requires CC: No

Free open source MCP. Requires language runtimes installed locally (Python, Node.js, etc.).

Agent Metadata

Pagination

none

Idempotent

Full

Retry Guidance

Not documented

Known Gotchas

⚠ Code runs with local user permissions — no sandboxing from filesystem or network
⚠ Language runtimes must be installed on host machine
⚠ Long-running or infinite loop code will block MCP — add execution timeouts
⚠ Community MCP from individual developer — limited security review
⚠ Stateful code execution not preserved between calls — each call is independent
⚠ Output size limits may truncate large results

Alternatives

mcp-sandbox csharp-runner

Full Evaluation Report

Detailed scoring breakdown, competitive positioning, security analysis, and improvement recommendations for Code Execution Mode MCP Server.

$99

API endpoint ↗ Agent guide ↗ Report inaccuracy

Scores are editorial opinions as of 2026-03-06.