MCP server that exposes skill operations and vault-provided tools for the sx CLI application
Single tool 'read_skill' has a basic description but lacks critical schema documentation. Input schema visible but minimal: only a 'name' parameter with type string and basic description. No output schema documented. Description is adequate (89 chars) but lacks context on when to use it, what it returns, or error cases. No parameter constraints (enums, patterns, length limits). Tool naming is acceptable (verb_noun form) but the server exposes only one capability, limiting assessment of composition patterns. No evidence of pagination, structured output documentation, or error recovery guidance.
Read a skill's full instructions and content. Returns the skill content as markdown with @file references resolved to absolute paths.
No output schema documented. LLMs cannot plan downstream operations or extract structured data from the response.
Input parameter 'name' lacks constraints. No enum, pattern, length limits, or format specification. LLMs may pass invalid skill names.
Tool description does not explain error cases or recovery paths. What happens if skill not found? Should LLM search first?
No pagination or result-limiting guidance. If skill content is large, response may exhaust context window.
Description does not clarify what '@file references resolved to absolute paths' means or how LLM should interpret the output.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | D | 56 | 2026-07-28+ | v2 |