openSenseMap MCP — citizen-science environmental sensor network (opensensemap.org)
Three well-defined tools with complete JSON schemas, clear descriptions, and good parameter documentation. Tool names follow verb_noun convention (opensensemap_*). Descriptions are detailed (150-400 chars) and include use cases, caveats about data quality, and examples. All parameters have type declarations and descriptions. However, output schemas are not documented, LLMs cannot predict response structure for planning. Error handling is minimal (only timeout/network errors visible). No input validation constraints (e.g., radius_km range 0.1-50 stated in description but not enforced via schema minima/maxima). Parameter descriptions could be more concise; some are verbose with repeated phenomenon hints.
Average one phenomenon across all citizen science sensors (openSenseMap / senseBox network) in an area — hyperlocal neighborhood-level temperature, PM2.5/PM10 air quality, humidity, pressure or noise from many independent community stations. Give either a bounding box or a center point + radius_km. Aggregates the latest reading per sensor over the recent window client-side (the network has no fast server-side aggregation). Phenomenon names on openSenseMap are mostly German (exact sensor titles): "Temperatur" (temperature °C), "rel. Luftfeuchte" (relative humidity %), "Luftdruck" (air pressure hPa), "PM2.5" and "PM10" (particulate matter µg/m³), "Beleuchtungsstärke" (illuminance lx), "UV-Intensität" (UV µW/cm²), "Lautstärke" (noise) — the name must match the sensor title exactly (e.g. "Temperatur", not "temperature"). Example: opensensemap_area_average({ phenomenon: "PM2.5", latitude: 52.52, longitude: 13.405, radius_km: 5 })
Get one openSenseMap citizen science sensor station (senseBox) by its box id — full sensor readout with latest value, unit, and measurement time per sensor (temperature, humidity, PM2.5/PM10 air quality, pressure, noise...), plus location, exposure (outdoor/indoor/mobile) and station metadata. Box ids come from opensensemap_nearby. Community-operated uncalibrated sensors. Example: opensensemap_box({ box_id: "65e8d93acbf5700007f920ca" })
Find citizen science sensor stations (senseBox, openSenseMap network) near a lat/lon and return their latest readings — hyperlocal temperature, humidity, air pressure, PM2.5/PM10 air quality, illuminance, UV, noise. Answers "sensor readings near me", "what does the local air quality sensor say". Community-operated uncalibrated hardware: quality varies, so cross-check outliers. Filter to one measurement type with `phenomenon` (Phenomenon names on openSenseMap are mostly German (exact sensor titles): "Temperatur" (temperature °C), "rel. Luftfeuchte" (relative humidity %), "Luftdruck" (air pressure hPa), "PM2.5" and "PM10" (particulate matter µg/m³), "Beleuchtungsstärke" (illuminance lx), "UV-Intensität" (UV µW/cm²), "Lautstärke" (noise)). Stations silent for more than 7 days are skipped unless include_stale=true. Example: opensensemap_nearby({ latitude: 52.52, longitude: 13.405, phenomenon: "PM2.5" })
Output schemas not documented. LLMs cannot predict response structure (fields, types, nesting) for downstream planning or data extraction. Breaks tool chaining.
Numeric parameter constraints (radius_km 0.1-50, limit 1-10, window_hours 1-24) stated in descriptions but not enforced in JSON Schema (missing minimum/maximum). LLMs may pass invalid values; validation happens at runtime only.
Error handling is minimal. Only network/timeout errors are caught. No guidance for common failures (invalid box_id, phenomenon not found, no stations in radius). LLMs cannot self-correct.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | D | 59 | 2026-07-28+ | v2 |
Phenomenon parameter in opensensemap_nearby accepts substring match (case-insensitive), but opensensemap_area_average requires exact match. Inconsistent behavior across tools with similar parameters.
PHENOMENON_HINT constant repeated in descriptions (opensensemap_nearby and opensensemap_area_average). Verbose; wastes tokens. Should be consolidated or moved to parameter-level descriptions only.