Loading ToolBench
MCP server integration package for the Daydreams AI framework, enabling tools and resources for agent deployment
This MCP server has severe gaps in definition quality. Only 1 tool is exposed ('add'), with a minimal schema and description. The tool definition is located in examples/mcp/mcp-server-example.mjs but the source code for that file is not provided, forcing assessment based on the JSON structure visible in the FEATURES list. The description 'Add two numbers' is generic and does not explain when to use this tool, what it returns, or any error cases. Parameters (a, b) have type definitions (number) but lack contextual descriptions beyond their names. The schema lacks documentation of the return type, pagination, or error handling. The server declares resources=true but provides no evidence of resource implementations in the provided source. This appears to be a documentation/example repository rather than a production MCP server, with the actual implementation details hidden.
Add two numbers
Tool description is too generic and does not meet LLM-optimization guidelines (50-200 chars, actionable context). 'Add two numbers' provides no context about when to use this tool, what it returns, or failure modes.
Parameters lack descriptions explaining their purpose or constraints. 'a' and 'b' have types but no contextual guidance for the LLM (e.g., range, precision, units). This violates the requirement that every parameter must have a non-empty description.
Return type and output schema are not documented. The LLM cannot infer what format the result takes (number, object, error structure). A documented output schema is required for proper tool composition.
No error handling documentation. The tool provides no recovery guidance for invalid inputs (e.g., non-numeric values) or overflow conditions. Error responses must tell the LLM what to do next.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-22 | F | 47 | 2026-07-28+ | v2 |
| 2026-03-09 | F | 34 | - | v1 |
Actual tool implementation (examples/mcp/mcp-server-example.mjs) is not provided in the source code, preventing verification of schema completeness, error handling, and actual behavior. Tool definition appears inferred rather than directly visible.
Tool name 'add' does not follow verb_noun naming convention. Consider 'calculate_sum' or 'add_numbers' to be more explicit about the action and make discovery clearer for LLMs.
Server declares resources=true but provides no implementation examples or documentation of available resources. This is either false advertising or incompletely implemented.