Loading ToolBench
Scoring was not performed
Missing output schema documentation. LLM cannot infer what fields/structure the analyze_plot response contains, forcing it to guess and potentially mis-parse results.
Tool description is minimal (47 chars) and lacks context. Does not explain WHEN to call analyze_plot vs other vision tools, WHAT analysis is performed (statistical interpretation, visual quality, data extraction?), or WHAT the LLM returns. Should be 100-200 chars with actionable guidance.
Optional parameter 'additional_instructions' is marked 'required: false' in the property definition but the key 'required' should be a boolean, not an array containing it. JSON Schema validation may reject this structure.
No error recovery guidance. If file not found, permission denied, or LLM timeout occurs, the tool should return actionable error messages guiding the agent on next steps (e.g., 'Image file not found at /path/to/file. Check the path and retry.').
Parameter 'image_path' requires 'full path' per description but no validation or format constraints are specified. Should include regex pattern (e.g., '^/.*\.png$') or format field to guide LLM input.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-22 | F | 0 | 2026-07-28+ | v2 |
| 2026-03-09 | F | 41 | - | v1 |