MCP server that provides access to Star Wars data via the SWAPI API, with tools for searching characters, retrieving planet/film information, and accessing internal character facts
Mixed quality. All 7 tools have names starting with action verbs and descriptions present. Input schemas are visible and use Zod validation with type definitions. However, descriptions lack depth for LLM optimization, most are under the recommended 50-200 character range (e.g., 'Get detailed planet info by its ID' is 34 chars). Parameter descriptions are minimal or absent in schema definitions. Three tools (prompt_*) are utility tools that expose prompt templates as tools, which is unconventional and mixes concerns. Output schemas are undocumented, LLMs cannot see what fields to expect in responses. Error handling exists but provides basic error text without recovery guidance. Tool composition is reasonable for a SWAPI wrapper, but prompt tools violate single-responsibility principle.
Get detailed film info by its ID
Get a private character fact from an internal system that is not in SWAPI
Get detailed planet info by its ID
Returns the analyze-character prompt template text for tools-only clients
Returns the compare-characters prompt template text for tools-only clients
Returns the explore-planet prompt template text for tools-only clients
Search for a Star Wars character by name
Tool descriptions are too short and lack LLM-optimization guidance. 'Get detailed planet info by its ID' (34 chars) does not explain WHEN to use it vs other tools or what the expected output structure is. Baseline production tools average 194 chars with clear context.
Output schemas are completely undocumented. Tools return JSON responses from SWAPI but LLMs have no declared structure showing what fields are in the response. This forces LLMs to parse unstructured output, wasting tokens and increasing hallucination risk.
Prompt tools (prompt_analyze_character, prompt_compare_characters, prompt_explore_planet) violate single-responsibility principle. These are not tools in the agent-interaction sense, they expose prompt templates as callable functions. This mixes the MCP tools layer with prompt engineering and creates confusion about when to invoke them.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-22 | D | 58 | 2026-07-28+ | v2 |
| 2026-03-09 | F | 37 | - | v1 |
Parameter descriptions in input schemas are minimal. 'Character name' and 'Planet ID' are self-evident from context but do not guide LLM behavior. No format constraints (e.g., 'ID must be numeric' or 'name minimum 2 chars') are documented in schema or description.
Error messages in error handler (makeToolError) are basic text without recovery guidance. Example: 'Error: SWAPI request failed: 404 Not Found' tells LLM nothing about what to do next, should suggest 'Try a different search term' or 'Available characters can be discovered via search_character'.