MCP metadata scoring and validation tool. Scores MCP server metadata against a rubric, captures live snapshots from stdio MCP servers, and computes deterministic diffs between snapshots.
Two fixture tools with minimal descriptions (both ~80 chars, baseline 194) and sparse parameter documentation. fixture_alpha has an enum constraint on 'kind' but no description for the parameter itself. fixture_beta's 'count' parameter lacks type bounds, range, or usage guidance. Neither tool description explains WHEN to use it, WHAT it returns, or WHY it exists beyond 'used by autobuilder capture ACs', insufficient for LLM selection. No output schemas documented. Tools appear to be test fixtures rather than production-grade agent tools, which explains the gaps, but they still fail baseline quality checks for parameter descriptions and output documentation.
Deterministic alpha fixture tool used by the autobuilder capture ACs.
Deterministic beta fixture tool used by the autobuilder capture ACs.
Parameter descriptions missing. 'kind' enum in fixture_alpha has no description; 'count' in fixture_beta has no description, type bounds, or usage guidance.
Output schemas not documented. Neither tool specifies what fields or structure it returns, forcing LLMs to guess downstream field names and types.
Tool descriptions too generic and context-dependent. Both describe purpose only as 'used by autobuilder capture ACs' without explaining WHAT the tool does, WHEN to call it, or WHAT it returns.
Numeric parameter 'count' in fixture_beta lacks bounds. No minimum/maximum specified; LLMs could pass negative, zero, or absurdly large values.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 40 | 2026-07-28+ | v2 |