MCP server for dynamically creating, managing, and executing other MCP servers from code templates in TypeScript, JavaScript, or Python
Server exposes 6 tools with basic schemas and descriptions, but critical gaps undermine production readiness. All tools have descriptions (10-80 chars, mostly adequate), and input schemas are present with type definitions. However, output schemas are completely undocumented, LLMs cannot predict what these tools return. Error handling is absent: no guidance on retryability, user-fixable vs fatal errors, or recovery paths. Tool composition is problematic: create_server and create_server_from_template overlap significantly; execute_tool accepts arbitrary args without validation. Security is a major concern: tools accept and execute arbitrary code (create_server, update_server) with no sandboxing, permission checks, or audit logging. Parameter descriptions lack constraints (e.g., no enum for language despite accepting only 'typescript'|'javascript'|'python'). No idempotency guarantees, no pagination for list results, no confirmation step before destructive operations (delete_server). Naming is verb-first and clear, but descriptions are too terse to guide LLM selection effectively.
Create a new MCP server from provided code in TypeScript, JavaScript, or Python
Create a new MCP server from a language template with optional custom code and dependencies
Stop and delete a running MCP server
Execute a tool on a running MCP server
List all available tools on a running MCP server
Update the code of a running MCP server
No output schemas documented. LLMs cannot predict return types, forcing them to guess what fields are available for downstream tool calls. This breaks tool chaining and forces discovery detours.
Arbitrary code execution without sandboxing, validation, or permission checks. create_server and update_server accept raw code strings and execute them directly. No audit logging, no dry-run, no confirmation step. Catastrophic security risk.
No error handling guidance. Tools provide no indication of retryability, user-fixable errors, or recovery paths. A failed execute_tool call returns nothing actionable to the LLM.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 43 | 2026-07-28+ | v2 |
Overlapping tool responsibilities. create_server and create_server_from_template both create servers; the distinction is unclear. LLMs will struggle to choose between them, wasting reasoning cycles.
Parameter descriptions lack constraints and validation rules. 'language' accepts enum values but description does not state the constraint. 'code' accepts arbitrary strings with no length limits or format guidance. 'args' in execute_tool is completely undocumented.