A source-available MCP server for HR and recruiting workflows. Provides AI-powered candidate scoring via Claude, with structured dimension scores, strengths, and gaps analysis.
Single tool 'score-candidate' has a well-structured schema with proper type definitions and reasonable descriptions. Input parameters are typed (string, object, integer) with constraints documented (character limits, weight bounds, positive integers). However, the tool lacks output schema documentation, the description mentions 'structured dimension scores' and 'overall fit score' but does not formally specify the response structure. Parameter descriptions are adequate (60-100 chars) but could be more LLM-optimized. No error handling guidance is visible. Tool naming is clear and verb-based ('score-candidate'), following the verb_noun pattern. The weights parameter is well-designed with nested properties and constraints, though the description could explicitly state the validation error message format.
Score a candidate's resume against a job description using Claude. Returns structured dimension scores (skills_match, experience, industry_background, education), strengths, gaps, and overall fit score.
Output schema not documented. Tool description mentions 'dimension scores', 'strengths', 'gaps', and 'overall fit score' but does not formally specify the response structure, field names, or types. LLMs cannot plan downstream operations without knowing what fields to extract.
No error handling guidance. Tool description does not explain what happens if resume_text or job_description is below minimum length, if weights do not sum to 1.0, or if the Claude API call fails. LLMs need recovery hints.
Parameter 'weights' lacks validation error messaging. Description states 'Must sum to 1.0' but does not explain what error the LLM will receive if it passes weights that sum to 0.95 or 1.05, preventing self-correction.
Tool description does not state whether the tool modifies state or is read-only. Although marked Risk: READ_ONLY in metadata, the description itself should explicitly say 'This tool does not modify any data; it is safe to call multiple times with the same inputs.'
Inferred effective spec: <=2025-11-25.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | D | 52 | <=2025-11-25 | v2 |
No guidance on when to use this tool vs. alternatives. Description does not explain prerequisites (e.g., 'Call this after you have both a resume and a job description') or when NOT to call it (e.g., if resume is too short).