Loading ToolBench
Advanced AI integration plugin that connects WooCommerce & WordPress with Model Context Protocol (MCP). Provides comprehensive AI-accessible interfaces to WooCommerce products, orders, categories, shipping, payments, and WordPress posts/pages through standardized tools, resources, and prompts. Enables AI assistants to seamlessly interact with your e-commerce data and content.
This WordPress/WooCommerce MCP server has 17 tools with READ_ONLY risk classification, but suffers from systematic gaps in schema documentation, parameter descriptions, and output specifications. Most tools lack complete input parameter schemas, parameter type information is sparse across the tool set, and descriptions are generic without strategic guidance for LLM selection. The server includes tool annotations (toolAnnotations=true) which is positive, but critical foundational quality issues prevent a higher score. Average per-tool score: 42/100.
Search and filter WordPress custom post types including with pagination
Get a WordPress custom post type by ID
Get the current logged-in user
Get WordPress general site settings
Get a WordPress media item details by ID
Get the actual file content (blob) of a WordPress media item
Get a WordPress page by ID
11 of 17 tools (wp_list_media, wp_pages_search, wp_get_page, wp_posts_search, wp_get_post, wp_list_categories, wp_list_tags, wp_get_general_settings, wp_get_user, wp_get_current_user, and others) have no visible input schema specification, violating the core requirement that parameters must be formally typed and constrained.
Most tool descriptions are bare one-liners (30-50 characters) that lack strategic guidance for LLM tool selection. Examples: 'List all WordPress post categories' (49 chars), 'Get a WordPress page by ID' (30 chars). Descriptions do not answer WHEN to use this tool vs similar tools, what the LLM should expect to get back, or any preconditions.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-22 | F | 45 | 2026-07-28+ | v2 |
| 2026-03-09 | F | 30 | - | v1 |
Get a WordPress post by ID
Get a WordPress user by ID
List all WordPress post categories
List WordPress media items with pagination and filtering
List all available WordPress custom post types
List all WordPress post tags
Search and filter WordPress pages with pagination
Search and filter WordPress posts with pagination
Search WordPress media items by title, caption, or description
Search and filter WordPress users with pagination
No output schemas documented. LLMs cannot know what fields to expect or what IDs to extract for downstream tool calls. E.g., does wp_posts_search return post_id, post_ID, or id? What fields are guaranteed? This breaks the tool chain pattern (pattern:tool-chain).
Parameter descriptions are missing for many tools. E.g., wp_users_search has 'context' as a required parameter with no description of what values are valid or what context controls.
No pagination guidance visible in descriptions. Tools like wp_posts_search and wp_pages_search likely support pagination but do not explicitly document limits, default page_size, or total_count return fields. This prevents LLMs from knowing when results are truncated.
No error handling documentation. Tools do not provide actionable error messages or recovery guidance. E.g., if wp_get_post fails with a 404, does it return 'not found', 'post_id not found', or an HTTP status? How should the LLM respond?
Some parameter names are under-specified. E.g., wp_cpt_search has 'author' (integer) but does not confirm it is author_id, author_user_id, or something else. wp_search_media has 'parent' (integer) but context is missing, parent post? parent category?
No enum constraints on string parameters. E.g., wp_cpt_search has 'status' parameter described as 'Filter by post status (publish, draft, pending, etc.)' but uses free-form string with no enum. LLMs will hallucinate invalid values like 'archived', 'pending_review', etc.