FastMCP 3.2+ server for USB logic analyzers via sigrok and simulator backends with protocol decoding and fleet webapp.
LogicAnalyzerMCP uses a 'portmanteau' pattern where 6 tools each accept an 'operation' enum parameter, collapsing multiple distinct operations into single tool definitions. While schemas are present with types and descriptions, this design violates the single-responsibility principle (pattern:tool). Tool names (la_device, la_configure, etc.) are generic and do not reflect specific actions, they read more like resource names than verb_noun action names. Descriptions are present but brief (50-100 chars), lacking context on when to use each operation or what state changes occur. Parameter descriptions exist but are minimal. No output schemas are documented. Error handling and recovery guidance are absent. The portmanteau pattern forces LLMs to reason about operation enums rather than selecting distinct, self-documenting tools.
Capture logic analyzer data. Operations: single, preview, export_csv, export_vcd, export_summary, last
Configure logic analyzer capture parameters. Operations: channels, sample_rate, get, simulator_pattern
Decode captured logic analyzer data using protocol decoders. Operations: list, uart, i2c, spi, last
Manage logic analyzer device connections and status. Operations: list, connect, disconnect, status, capabilities, backends
Get help and information about logic analyzer usage. Operations: discover, tool_help, status, quickstart, faq, hardware_guide
Configure trigger conditions for capture. Operations: set, get
Portmanteau anti-pattern: all 6 tools collapse multiple operations into a single 'operation' enum parameter. This violates single-responsibility principle and forces LLMs to reason about operation strings rather than selecting distinct, self-documenting tools. E.g., 'la_device' handles list, connect, disconnect, status, capabilities, backends, six separate concerns.
Tool names are resource-centric (la_device, la_configure) rather than action-centric (list_devices, connect_device, configure_channels). LLMs infer intent from verb_noun naming; generic names like 'la_device' do not signal what action occurs.
No output schemas documented. LLMs cannot plan downstream tool calls or extract required fields (e.g., device_id from list operation, or decoded data structure from la_decode). Responses are unstructured.
Inferred effective spec: <=2025-11-25.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 45 | <=2025-11-25 | v2 |
Descriptions are minimal (50-100 chars) and lack context on state changes, prerequisites, or when to use each operation. E.g., 'Manage logic analyzer device connections and status' does not explain that 'connect' is destructive or that 'list' is read-only.
No error handling or recovery guidance. If a device connection fails or a capture times out, LLMs receive no actionable error message or retry strategy.