MCP Server for executing Python code in an isolated REPL namespace with captured stdout/stderr
Single tool with minimal definition quality. Tool name 'run_python_code' is generic and lacks action clarity (verb_noun pattern weak, 'execute_python_code' would be stronger). Description is present but extremely brief (11 words, 89 chars, below the 10 - 1024 char baseline and lacks WHEN/WHY context). Input schema exists but parameter 'code' has NO description field, violates critical rule that every parameter needs a description. No output schema documented. No error handling guidance. No security considerations for arbitrary code execution (WRITE risk). Tool definition is directly visible in mcp_server.py, so not capped at 50 for inference.
Execute Python code in an isolated REPL namespace with captured stdout.
Parameter 'code' lacks description. Schema shows type:string but no guidance on what code is valid, constraints, or examples. LLMs cannot infer whether to pass raw Python, imports, or multi-line scripts.
Tool description is 89 characters and lacks WHEN/WHY context. Does not explain when to use this vs other tools, what happens on error, or what the output structure contains. Baseline is 194 chars for production tools.
No output schema documented. Response wraps result in JSON string ('json.dumps(res)') rather than structured fields. LLMs cannot parse stdout, stderr, eval_result, error, elapsed_sec, active_variables without reverse-engineering the JSON.
No error handling guidance. Tool executes arbitrary Python code (WRITE risk) with no validation, sandboxing limits, or recovery hints. If code raises an exception, response includes raw error message but no actionable next steps for LLM.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 37 | 2026-07-28+ | v2 |
Generic tool name 'run_python_code' does not follow verb_noun convention clearly. 'execute_python_code' or 'eval_python_expression' would be more specific. Current name conflates execution modes (eval vs exec) without clarity.