MCP server for accessing BCRP (Banco Central de Reserva del Perú) time series data with search and retrieval capabilities
The server defines 3 tools with clear action-verb naming (search_*, get_*) and reasonable descriptions (120-200 chars). All tools are READ_ONLY, reducing risk. However, parameter descriptions lack depth and format constraints. Input schemas are present but minimal, no enums, ranges, or validation patterns. Output schemas are not documented, forcing LLMs to infer structure. Error handling is absent; no recovery guidance or actionable error messages visible. The descriptions mention BCRP database and date format hints ('2020-1' for monthly, '2020-1-1' for daily) but lack formal constraints. Overall, definitions are functional but lack the rigor expected of production-grade tools.
Get the data for a specific time series within a date range. This function retrieves time series data from the BCRP (Banco Central de Reserva del Perú) database for a specific time series code within the specified date range. The data is returned as a list of lists with dates formatted as 'YYYY-MM-DD'.
Search for time series within a specific group from the BCRP database. This function retrieves time series metadata from the BCRP (Banco Central de Reserva del Perú) database and filters it to find all time series that belong to a specific group. It returns a list of dictionaries containing the code and name of each matching time series.
Search for time series groups using one or multiple keywords. This function searches the BCRP (Banco Central de Reserva del Perú) database for time series groups that match the provided keywords. It returns a list of unique time series group names that contain or relate to the search terms.
Output schemas not documented. Tools return data structures (lists of dicts, date-value pairs) but LLMs cannot see the field names, types, or nesting. This forces inference and increases hallucination risk.
Date format parameters lack formal constraints. Descriptions say 'Format should be 2020-1 for monthly or 2020-1-1 for daily' (natural language), but no regex pattern or enum in schema. LLMs may infer wrong formats.
No error handling guidance. If a time series code is invalid, if dates are out of range, or if the BCRP API is unavailable, there is no recovery hint for the LLM. Errors will deadend agent execution.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-22 | D | 50 | 2026-07-28+ | v2 |
| 2026-03-09 | D | 50 | - | v1 |
Parameter 'time_series_group' in search_time_series_by_group is vague. Description says 'should match or be contained within' but does not clarify exact matching rules, case sensitivity, or partial matching behavior.
No pagination or result limiting visible in get_time_series_data. If a time series spans decades with daily data, returning thousands of rows will blow the context window. Schema should support limit/offset and return a count.