Loading ToolBench
MCP server for GoodData analytics platform integration, enabling visualization search, export, and scheduling capabilities
Three tools with basic schemas and descriptions, but significant gaps in parameter documentation and error handling. visualization_search and visualization_png_export have minimal descriptions (under 50 chars). visualization_schedule has better description but lacks parameter validation guidance. No output schemas documented. Parameters lack type constraints (e.g., email format, cron validation). No error recovery guidance. Missing idempotency hints for the write operation. Naming is acceptable (verb_noun pattern) but descriptions are too terse to guide LLM selection effectively.
Export a visualization to image
Schedule a visualization export over email. Cron format must be: SECOND MINUTE HOUR DAY-OF-MONTH MONTH DAY-OF-WEEK For example, every Monday at 9am would be: 0 0 9 * * MON
Find relevant visualizations by search term
Parameter descriptions missing or incomplete. 'visualizationId' and 'email' lack format/constraint guidance. 'cron' parameter has example but no validation rules (e.g., 'must be valid cron format with 6 fields'). LLMs cannot infer valid input without explicit constraints.
Tool descriptions too brief (10-25 chars). 'Find relevant visualizations by search term' and 'Export a visualization to image' lack context for LLM selection. No guidance on when to use each tool vs alternatives, prerequisites, or return structure.
No output schemas documented. LLMs cannot plan downstream calls or extract required fields. visualization_search returns text results; visualization_png_export returns base64 images; visualization_schedule returns automation ID, none explicitly documented.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 43 | 2026-07-28+ | v2 |
No error handling guidance. visualization_png_export polls with MAX_POLL_ATTEMPTS=10 but returns generic 'Export failed' on timeout. No guidance to LLM: retry? Try different visualization? Check permissions? visualization_schedule creates automation with no validation of cron format before API call.
visualization_schedule is a write operation but lacks idempotency hint and confirmation pattern. Creates automation with timestamp-based ID (automation-${Date.now()}), repeated calls with same params create duplicate automations. No dry-run or confirmation step.