Model Context Protocol server for BMLT (Basic Meeting List Tool) API with intelligent geocoding and rate limiting
This BMLT server has solid parameter schemas with good type definitions and reasonable descriptions across most tools. However, it suffers from several critical gaps: (1) Tool descriptions are generally brief but adequate (60-200 chars), though some lack actionable context for LLM selection. (2) No output schemas are documented, we can see input schemas but no description of what these tools return, forcing LLMs to guess at response structure. (3) Parameter descriptions are present but inconsistent in depth, some explain context (e.g., 'Address searches are automatically geocoded'), others are minimal. (4) No error handling guidance is visible, no recovery strategies, no error classification, no actionable error messages. (5) The tools are read-only with no destructive operations, which is good for safety, but the descriptions do not explicitly state this. (6) Parameter dependencies are implicit (e.g., geo_width vs geo_width_km, lat_val + long_val together) and should be documented. Overall, this is a data-retrieval API wrapper with decent structure but lacks the documentation depth needed for production LLM agent work. Per-tool scores average 62.
Retrieve meeting changes within a date range
Get geographic coverage area information
Get list of available field keys
Get specific field values for a given field key
Retrieve available meeting formats from BMLT server
Export meeting data in NAWS format (CSV)
Get server information including version and configuration
No documented output schemas. All 9 tools lack descriptions of what they return (response structure, field names, types). LLMs cannot plan downstream tool calls or extract required fields without knowing the response shape.
No error handling guidance. Tool descriptions do not explain what errors might occur, whether they are retryable, or how to recover. E.g., 'meeting_ids' can accept negative values to exclude, what happens if an invalid ID is passed? What does the API return?
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-22 | D | 55 | 2026-07-28+ | v2 |
| 2026-03-09 | F | 45 | - | v1 |
Get list of service bodies from BMLT server
Search for meetings in the BMLT database with various filtering options including published status filtering. For location-based searches, you can use either address strings (which will be automatically geocoded) or specific lat/long coordinates for more reliable results.
Implicit parameter dependencies not documented. bmlt_search_meetings accepts both 'SearchStringRadius' and separate 'lat_val'/'long_val' coordinates. The interaction is unclear: can both be used? What if they conflict? Dependencies like 'geo_width' vs 'geo_width_km' (miles vs kilometers) could cause silent misuse if not explicitly documented.
Generic/minimal descriptions for 6 tools. 'Get list of service bodies from BMLT server' (48 chars) and 'Get specific field values for a given field key' (46 chars) lack context for when/why an LLM should call them vs alternatives. Baseline is 194 chars; these are significantly below and do not explain use cases or dependencies.
No pagination guidance documented. bmlt_search_meetings accepts 'page_size' and 'page_num' but the description does not state maximum page size, total result count limits, or when pagination is required. Without this, agents may request enormous pages or miss multi-page results.
Tool descriptions do not explicitly state read-only nature. No tool description says 'This is read-only' or 'No data is modified.' LLMs cannot distinguish safe-to-retry from destructive operations without this clarity. Even though all tools are safe, documentation should make this explicit for agent planning.