IMA Knowledge Base Management

The Problem

Engineers integrating IMA knowledge bases often deal with OpenAPI calls where field names, kb_id, folder_id, and media_id are easy to mix up, while writes are hard to reverse. This skill turns common IMA knowledge base management operations into a concrete workflow: locate the knowledge base, browse folders, import content, and maintain tags and permissions, reducing trial-and-error from memory-based API calls.

How It Works

• Search and locate: use search_knowledge_base to get the knowledge base ID and role, get_knowledge_list to browse folder structure, and search_knowledge to locate files or folders by name.
• Content onboarding: supports upload_file.py for files, import_urls for web links, and add_knowledge for notes, useful for consolidating scattered materials into one knowledge base.
• Structure and tags: create folders, rename, pin, and move files, while managing tags with tag_add, tag_remove, tag_rename, and tag_delete.
• Permissions: control access through join_knowledge, update_knowledge_base_permission, and update_knowledge_access_status.

Boundaries

• The current API does not support deleting knowledge bases, folders, files, or uploaded content; deletion requests should be handled in the IMA client.
• Writes are irreversible, especially tag_delete and tag_rename; preview affected files with tag_list or get_knowledge_list(tags=[...]) before executing.
• Content search and reading are better handled by search / fetch; this skill focuses on resolving kb_id, browsing metadata, importing/exporting, and managing permissions.
• All returned IDs must be passed exactly as-is, without trimming, prefixing, or reformatting.