Shared MCP server factory for Upvote.club Public API — all 7 backend endpoints as MCP tools for social media promotion task management
Strong tool naming (verb_noun pattern), comprehensive descriptions (avg 180 chars), and detailed input schemas with enums and constraints. All 6 tools have descriptions and most parameters are typed. However, output schemas are not formally documented in the code, responses are inferred from API reference text rather than explicit JSON Schema definitions. Error handling is present but generic (toolResult with isError flag). Missing per-tool output schema documentation and some parameter descriptions lack actionable constraint details.
Create a promotion task on Upvote.club. SPENDS points (price × actions_required, plan discount) and 1 daily slot. Call get_api_reference for examples and errors. Auto-detects platform from URL when social_network_code omitted. GitHub/Product Hunt min 16 pts/action; meaningful COMMENT min 64 pts/action.
Delete a task you created. Refunds points for incomplete actions and restores 1 available_tasks slot.
Get your account balance, plan, available task slots, and API key info.
Full Upvote.club Public API reference for MCP: all tool inputs/outputs, request examples (single-action, MULTI_ACTION, meaningful comments), minimum prices per platform/action, billing rules, and complete error catalog. Call before create_task if unsure about fields or pricing.
Track progress for one or more tasks you created.
List all social platforms supported by Upvote.club with their available promotion actions, minimum price per action (in points), and example URLs. Call this first if you are unsure which platform or action applies to the link the user wants to promote.
Output schemas not formally documented. Responses are returned as JSON text via toolResult() with no explicit schema definitions. LLMs cannot predict response structure for planning downstream calls.
get_task_status and get_account_info descriptions are generic (70 chars). Missing context on when to call, what data is returned, and how to use results for downstream decisions.
create_task schema is complex with nested conditional logic (meaningful_comment vs meaningful_comments, single-action vs MULTI_ACTION). Dependencies between parameters not explicitly documented in descriptions.
Inferred effective spec: <=2025-11-25.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | C | 64 | <=2025-11-25 | v2 |
Error responses use generic toolResult(payload, isError=true) pattern. No actionable recovery guidance (e.g., 'API key not configured, set UPVOTE_API_KEY env var'). Some errors logged to stderr but not returned to LLM.
STDIO transport only. Server is not remotely accessible and cannot be used by hosted MCP clients. Limits deployment flexibility.