The Agent Skills Directory

[SAFE]: The skill implements a secure-by-default workflow, requiring the agent to explain planned index configurations and obtain explicit user approval before executing the create-index tool.\n- [SAFE]: The instructions mandate an 'inspect first' approach using tools like collection-schema and collection-indexes to validate the environment and use case before providing recommendations.\n- [SAFE]: The skill includes defensive version checks for advanced MongoDB features (like $rankFusion and $scoreFusion), ensuring compatibility and preventing execution on unsupported clusters.\n- [SAFE]: Technical references for advanced search operators (e.g., queryString) include explicit warnings to validate and sanitize user input to prevent injection attacks.

mongodb-search-and-ai