An MCP server that integrates GitHub API access with an LLM interface, allowing users to query their GitHub repositories through natural language
This server has critical definition quality gaps. The single tool 'get_repos' lacks a proper input schema entirely (no parameters defined in the code), has a minimal 30-character description, and no documented output schema. The tool is not registered via standard MCP mechanisms, it's a Flask HTTP endpoint. The server does not implement MCP protocol at all; it's a custom HTTP API masquerading as an MCP server. No tool annotations, no error guidance, no pagination support despite returning potentially large lists.
Fetches Names of all Repositories.
No input schema defined for get_repos. Tool accepts no documented parameters, making it impossible for LLMs to understand what inputs it accepts or how to invoke it correctly.
Description is only 30 characters ('Fetches Names of all Repositories.').
No output schema documented. LLMs cannot plan downstream calls or extract data when the response structure is unknown. The code returns a newline-delimited string of repo names, not structured JSON.
No pagination support. If a user has 100+ repositories, the tool returns all names concatenated as a single string, wasting tokens and risking context window exhaustion. No limit parameter, no cursor, no total count.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 24 | 2026-07-28+ | v2 |
No error handling or recovery guidance. If auth_data['username'] is None (user not authenticated), the tool silently fails with a 401. LLMs receive no actionable error message telling them to authenticate first.