Loading ToolBench
Lyearn MCP server with auto authentication support
Single tool with basic schema and minimal descriptions. Tool name 'create_action_item' follows verb_noun convention (+), but description is generic (18 chars, below 20-char threshold). Input schema present with type definitions, but parameter descriptions are extremely terse ('Title of Action Item', 'Status of Action Item', 22 and 24 chars respectively, below baseline 72 chars). No output schema documented. No error handling guidance. Tool does not validate status enum values, inviting hallucinated statuses. Missing idempotency guidance for a write operation. No pagination, no chaining IDs, no recovery hints.
Create a new Action Item
Tool description is 18 characters ('Create a new Action Item'), below 20-char minimum. Lacks context for LLM selection: does not state WHEN to use it, WHAT it returns, or that it modifies state.
Parameter 'status' accepts free-form strings with no enum constraint. LLM will hallucinate invalid statuses (e.g. 'pending', 'done', 'archived') instead of valid ones. No description of allowed values.
No output schema documented. LLM cannot infer what fields the response contains (action_item_id? created_at? url?). Breaks tool chaining, downstream tools cannot reference the created item.
Parameter descriptions are 22 - 24 characters (baseline 72 chars). 'Title of Action Item' and 'Status of Action Item' do not explain format, constraints, or examples. LLM cannot validate input before calling.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 46 | 2026-07-28+ | v2 |
No error handling guidance. If status is invalid or title is empty, error response is generic ('Invalid arguments provided'). LLM cannot self-correct or know what to try next.