Unofficial third-party Python client and MCP server for the RepeaterBook.com API, exposing RepeaterBook lookup tools to agents.
RepeaterBook MCP has four well-named tools with clear purposes and complete input schemas. All tool names follow verb_noun convention (search_repeaters, get_repeater, sync_repeaters, clear_local_data). All four tools have descriptions (range 90-178 chars, within the 10-1024 baseline). All parameters have type definitions and descriptions. However, output schemas are NOT documented, we can infer responses exist but cannot see the return type definitions in the source code provided. Error handling is not visible. Tool descriptions lack dependency hints or recovery guidance. Parameter descriptions are present but generic; some could be more instructive about constraints and formats. The server demonstrates solid foundational quality but lacks depth in output documentation and error guidance expected of production tools.
Empty the local repeater store and clear the API response cache. Both must be cleared together to ensure the next sync is a genuine download.
Get details of a single repeater by its source ID (state_id:repeater_id format).
Search for repeaters by location and filters (distance radius, bands, modes, status, use). Returns repeaters sorted by distance.
Download repeaters for a geographic scope from RepeaterBook and store them locally. Returns sync result with counts and truncation status.
Output schemas not documented. Tool definitions show input schemas clearly but return types are entirely absent from the source code audit. LLMs cannot plan downstream actions or extract return data without knowing field names and types.
Error handling and recovery guidance missing. No evidence of error classification (retryable vs user-fixable vs fatal), actionable error messages, or recovery hints. Tool descriptions do not explain what to do if a search fails, API is down, or credentials are invalid.
Destructive tool (clear_local_data) lacks confirmation or dry-run pattern. No evidence of a confirmation step or a way to preview the destructive action before execution. Agents could accidentally clear data without safeguards.
Inferred effective spec: <=2025-11-25.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | C | 66 | <=2025-11-25 | v2 |
Parameter descriptions generic; lack constraint details. E.g., 'radius_km' has no minimum/maximum bounds documented; 'country' says 'e.g. United States' instead of constraining to valid country names; 'modes' enums are present but not all parameters specify valid ranges or formats.
Pagination not documented. search_repeaters returns repeaters but no limit, offset, or pagination parameters visible. If results exceed context window, the tool response could blow context or silently truncate.