Related Servers

Most tools target a clearly distinct resource+action (ingest_html vs ingest_url by input type, search_images vs upload_images, render_preview vs publish_page), and the long descriptions actively draw boundaries. The confusable spots are list_pages vs find_pages (find_pages with no filters duplicates list_pages' behavior) and the four write paths create_page / update_page / add_section / patch_page, where an agent must read carefully to pick full-replace versus server-side append versus id-level patch. validate_page also overlaps conceptually with the dry_run validation built into every save tool.

Almost everything is snake_case verb_noun (get_page, list_pages, create_page, update_page, publish_page, patch_page, search_images, upload_images, validate_page). Two deviations: the bare noun 'layout' breaks the verb pattern, and the 'new_' prefix (new_page_skeleton, new_element) sits alongside 'create_' (create_page), which slightly muddies whether 'new' means create-something or return-a-template.

23 tools is at the heavy end, and the surface spans several sub-domains at once: page CRUD, element reference docs (get_page_schema, list_elements, get_element, get_generation_guide), layout/compute helpers, asset search/upload/icon tools, ingestion, and rendering. Each tool is individually defensible for a page builder, but four of them are static reference/documentation lookups and the create/update/patch/add_section quartet is more save-paths than strictly necessary.

The build lifecycle is well covered end to end: skeleton, element factory, layout computation, validation, create/read/list/find/update/patch/append/publish, plus asset ingestion, image upload/search and preview rendering. Notable gaps remain: there is no delete_page or unpublish/archive operation, and no way to list or manage already-uploaded media in the organization collection (only upload and external Pexels search), so an agent cannot clean up or reuse prior uploads.