hybrid-search-eval-author
Pass
Audited by Gen Agent Trust Hub on Aug 12, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill uses established retrieval evaluation metrics and provides code examples for interacting with official APIs of well-known services (Elasticsearch, OpenSearch, Cohere, Anthropic).
- [EXTERNAL_DOWNLOADS]: The skill references and provides links to official documentation for Elasticsearch, OpenSearch, and Cohere. These are well-known technology providers and the links target their official documentation domains (elastic.co, opensearch.org, docs.cohere.com).
- [COMMAND_EXECUTION]: The skill demonstrates standard API interactions using official client libraries (e.g.,
anthropic,cohere) and HTTP requests to search engine endpoints. All operations are aligned with the skill's primary purpose of search relevance evaluation. - [DATA_EXPOSURE]: The instructions recommend standard practices for managing query and document data for evaluation. It includes a section on LLM-assisted labeling which involves sending text fragments to the Anthropic API, which is a standard use case for search labeling.
Audit Metadata