API hh.ru для ИИ-ассистентов: вакансии, отклики и приглашения, резюме, справочники, статистика зарплат
Single tool 'hh_get_me' has minimal inline definition. Tool is registered via schema_mcp_core.tools.register_generic_tools() and register_cabinet_tools(), which means tool definitions are generated from external YAML catalogs (endpoints.yaml, entities.yaml) not visible in the provided source. The server.py file shows only high-level registration calls without inline schema or parameter documentation. No input validation, error handling guidance, or output schema documentation visible in the code.
Get current user information from hh.ru API
Tool definitions sourced from external YAML files (endpoints.yaml, entities.yaml) not provided in source code. Cannot verify schemas, parameter types, descriptions, or output structure.
No input schema visible for hh_get_me in provided source. Tool registration delegates to schema_mcp_core library functions, making schema validation impossible from static analysis.
No error handling guidance visible. Tool description in server.py is minimal ('Get current user information from hh.ru API'). No documentation of what happens on auth failure, network error, or invalid token. Per pattern:recovery-guide, errors must tell the LLM what to do next.
Credentials (HH_TOKEN, HH_APP_NAME) passed via environment variables but no documentation of how the tool handles missing or invalid credentials. No guidance on token acquisition or error messages for auth failures.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 39 | 2026-07-28+ | v2 |
No output schema documented. Tool returns 'current user information' but structure (fields, types, nested objects) is not specified in visible code. LLM cannot plan downstream calls without knowing what fields are available.