Loading ToolBench
MCP Server for Kalman Filter 1D State Estimator Skill - performs recursive Predict-Update cycles for continuous latent state estimation under Gaussian noise
Single tool with minimal definition quality. Tool name 'filter_measurement' lacks action verb clarity (should be 'apply_filter' or 'estimate_state'). Description is adequate (89 chars, within 10-1024 range) but lacks WHEN/WHY guidance. Input schema present but parameters lack descriptions entirely, critical violation of pattern:tool-description. No output schema documented. No error handling guidance. No parameter constraints (measurement and control are unbounded floats). Tool does not follow verb_noun naming convention strongly enough.
Performs a single Kalman filter step: predict phase with optional control input, then measurement update phase. Returns estimated state and variance.
Input parameters lack descriptions entirely. 'measurement' and 'control' parameters have no descriptions explaining what values are valid, expected ranges, or units.
No output schema documented. Response returns 'estimated_state' and 'variance' but LLM has no formal schema defining types, ranges, or meaning of these fields.
Tool name 'filter_measurement' is weak. Does not start with clear action verb. 'apply_kalman_filter' or 'estimate_state' would be clearer for LLM intent inference.
No parameter constraints. 'measurement' and 'control' accept unbounded floats with no min/max, units, or valid range guidance. LLM may pass absurd values.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 35 | 2026-07-28+ | v2 |
No error handling or recovery guidance. Tool returns generic error dict on unknown method but provides no actionable guidance for LLM on what went wrong or how to recover.