A production-grade MCP server exposing Derek Huynen's public professional profile over stdio and Streamable HTTP.
Six tools with clear verb-noun naming (get_*, search_*) and consistent descriptions (10-50 chars). All tools have input schemas with Zod validation and type definitions. However, parameter descriptions are minimal (e.g., 'Filter by skill keyword' lacks format/constraint details), and output schemas are not documented. No error handling guidance, no pagination support despite potential for large result sets, and no tool annotations (readOnlyHint). Tools are read-only and well-scoped, but descriptions could be more LLM-optimized per pattern:tool-description baseline (194 chars avg).
How to reach Derek Huynen (email and profile links).
Derek's roles and engagements, optionally filtered.
Derek Huynen's identity, title, summary, and links.
Derek's public portfolio projects.
Derek's skills grouped by category.
Keyword search across roles and projects.
Output schemas not documented. LLMs cannot infer what fields to expect from tool responses, forcing them to guess at downstream field names and risk chaining failures.
Parameter descriptions lack format/constraint details. 'Filter by skill keyword, e.g. "RAG"' does not specify case sensitivity, max length, or whether partial matches are supported.
No pagination support. If get_projects or get_experience return large result sets, responses could exceed context windows. No limit parameter, no cursor, no total count documented.
Inferred effective spec: 2026-07-28+.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | C | 60 | 2026-07-28+ | v2 |
No tool annotations. Tools lack readOnlyHint, destructiveHint, or idempotentHint metadata. While all tools are read-only, explicit annotations per MCP spec 2026-07-28 would improve client UX and safety.
No error handling guidance. Tools return JSON results but do not document what happens on invalid input (e.g., invalid skill filter, malformed query). No recovery hints for LLMs.