Loading ToolBench
Visual Paradigm 17.0 plug-in exposing the open project over the Model Context Protocol.
Server has 2 tools with complete input schemas and descriptions. Echo is a test utility with minimal value. GetDiagramByUrl has a well-structured schema with enum constraint on 'detail' parameter and clear descriptions. However, output schemas are not documented, LLMs cannot predict the structure of returned diagram JSON. Parameter descriptions are adequate but lack format/constraint details (e.g., vpp_url format expectations). No error handling guidance provided. Missing tool annotations (readOnlyHint, idempotentHint) despite both tools being read-only operations.
Echo the given message back. Exists so the MCP transport can be tested without Visual Paradigm.
Read a diagram from the Visual Paradigm project that is currently open, and return it as a JSON graph. Every node and edge carries both its model data (name, type, documentation, stereotypes, tagged values, attributes and operations) and its exact formatting (fill colour, line colour, line style and weight, font colour, bold and italic, caption placement, bounds). Colours are reported as they are, for example "#FF0000"; the meaning of a colour is not interpreted here.
Output schemas not documented. LLMs cannot predict the structure of diagram JSON returned by get_diagram_by_url. Pattern requires documenting return types so agents can plan downstream operations.
Missing tool annotations. Both tools are read-only but lack readOnlyHint annotation. This prevents agents from understanding which tools are safe to call speculatively vs. which require confirmation.
No error handling guidance. Tools do not document what errors can occur, whether they are retryable, or what the LLM should do next. E.g., get_diagram_by_url should explain: 'Diagram not found, verify vpp_url format or check if project is open.'
Parameter vpp_url lacks format specification. Description says 'e.g. MOM.vpp://diagram/YRvbHuaFYFAEMEOa' but does not state whether bare IDs are truly accepted or what the regex pattern is. LLMs may pass invalid formats.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | C | 63 | 2026-07-28+ | v2 |
Echo tool has minimal utility and no clear use case beyond transport testing. Consider removing or documenting when agents should call it.