Simulated IoT sensor fleet exposed as an MCP server via FastMCP
FastMCP IoT Sensors has 8 tools with complete input schemas and descriptions, but quality is uneven. Tool names follow verb_noun convention (get_, set_, list_, acknowledge_). Descriptions range from adequate (40-80 chars) to good (100+ chars). However, parameter descriptions are sparse or missing context. Output schemas are not documented in the visible code. Error handling is absent, no recovery guidance, no validation messages. The server lacks tool annotations (readOnlyHint, destructiveHint) despite clear risk classifications. Composition is reasonable: tools are single-purpose and chainable (e.g., list_sensors → get_sensor_reading). Overall, definitions are functional but lack LLM-optimization and production rigor.
Mark an alert as handled/acknowledged.
Acknowledge multiple unacknowledged alerts at once, confirming scope with the user first if there's more than one active alert.
List unacknowledged alerts, most recent first.
Return the time-series of readings for a sensor over the past since_minutes minutes.
Return the latest reading for a single sensor by id.
Return all sensors with their current status (id, name, type, zone, last value, last updated).
Configure the alert min/max bounds for a sensor.
No output schemas documented. Tools return structured objects (SensorStatus, ReadingPoint models visible in models.py) but MCP tool definitions do not declare return types. LLMs cannot plan downstream calls or extract fields without explicit schema.
Missing tool annotations. Tools have explicit risk classifications (WRITE, READ_ONLY) but FastMCP tool definitions lack readOnlyHint and destructiveHint annotations. Agents cannot distinguish safe from destructive calls without explicit hints.
No error handling or recovery guidance. Tools have no documented error cases, validation rules, or recovery paths. E.g., set_alert_threshold accepts min/max but does not document what happens if min >= max, or guide the LLM to retry with corrected values.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | D | 58 | 2026-07-28+ | v2 |
Gather current sensor values and active alerts for a zone, then ask the client's LLM to produce a short natural-language health summary.
Parameter descriptions lack context and constraints. 'sensor_id' is described as 'The sensor ID to configure' but does not explain valid range, what happens if ID does not exist, or how to discover valid IDs. 'since_minutes' has min/max in schema but description does not explain the constraint rationale.
bulk_acknowledge_alerts has empty input schema ({}). Tool description mentions 'confirming scope with the user first' but no parameters exist to control this behavior or specify which alerts to acknowledge. Unclear how the tool determines scope without input.