Clusters GeoJSON points using DBSCAN algorithm and returns them as circles (polygons)
Single tool 'cluster_points' has a clear verb-noun name and acceptable description (89 chars). Input schema is present with three parameters (geojson, eps, minPts), all with type definitions and descriptions. However, the schema lacks critical details: no minLength/maxLength for geojson string, no minimum/maximum bounds for numeric eps/minPts parameters, no enum constraints, and no documented output schema. The description does not explain what the tool returns (GeoJSON FeatureCollection? Polygon coordinates? Cluster metadata?). No error handling guidance, no recovery hints, and no mention of prerequisites or when to use this tool. Parameter descriptions are minimal (e.g., 'Maximum distance between two points to be considered as in the same neighborhood (meters)' is adequate but lacks range constraints). The tool is read-only (good), but lacks idempotent/destructive annotations. Overall, the tool is functional but underspecified for LLM-driven agent use.
Clusters GeoJSON points and returns them as circles (polygons).
Output schema not documented. Tool description does not specify what cluster_points returns (structure, fields, format). LLMs cannot plan downstream operations or extract cluster IDs/metadata without knowing the response shape.
Numeric parameters lack bounds. 'eps' and 'minPts' have no min/max constraints documented. LLMs may pass absurd values (eps=999999, minPts=1000000) that break clustering or cause timeouts.
No error handling guidance. Tool description does not explain what happens on invalid GeoJSON, empty point sets, or parameter violations. LLMs have no recovery path if clustering fails.
GeoJSON parameter accepts free-form string with no validation hints. Description should specify expected structure (FeatureCollection with Point features), coordinate system (WGS84?), and size limits to prevent malformed input.
Inferred effective spec: <=2025-11-25.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 49 | <=2025-11-25 | v2 |
Tool description lacks context for LLM selection. Does not explain when to use cluster_points vs other spatial tools, what clustering algorithm is used (DBSCAN?), or what the output represents (cluster polygons, centroids, or both?).