MCP server for Minecraft (Spigot/Paper) that exposes server management tools and resources via the Model Context Protocol
No tool descriptions explain error recovery, prerequisites, or when to use each tool vs alternatives. Naming is acceptable (verb_noun pattern) but descriptions are generic. No output schemas documented. Error handling is absent, no guidance on retryability, user-fixable vs fatal errors, or recovery steps.
Write or append raw bytes to a file under the server root (creates parent directories). Content is Base64-encoded in the contentBase64 argument.
Run a command as the console (per-tool commandWhitelist/commandBlacklist in config)
Read the tail of logs/latest.log (server directory). Use tools.*.maxLogLines, maxLogBytes in config for defaults.
fs_write and server_command parameters lack descriptions. 'path', 'contentBase64', 'append', 'command' have no explanation of format, constraints, or expected values. LLMs cannot infer whether 'path' is absolute or relative, or what 'contentBase64' encoding means without explicit docs.
No output schemas documented for any tool. Callers do not know what fields to expect (e.g., does server_logs return {lines: string[], totalLines: int} or {content: string}?). This forces LLMs to guess and breaks downstream tool chaining.
No error handling guidance. fs_write (WRITE risk) and server_command (IRREVERSIBLE risk) do not document what happens on failure, whether errors are retryable, or how to recover. Agents cannot distinguish 'disk full' (retry later) from 'permission denied' (ask user).
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 45 | 2026-07-28+ | v2 |
server_command description mentions 'commandWhitelist/commandBlacklist in config' but does not explain what happens if a command is blocked. Does the tool reject it? Silently ignore it? Return an error? Ambiguity invites misuse.
fs_write description does not warn about path traversal risks or clarify that 'server root' is a security boundary. No mention of what happens if parent directories cannot be created (permissions, disk space). Agents may attempt unsafe paths.