Loading ToolBench
MCP server that exposes documentation content as tools and resources. Provides full-text search, page reading, navigation, and tag-based filtering across markdown documentation.
Six tools with complete input schemas and descriptions. All tools follow verb_noun naming (search_docs, read_page, list_pages, get_table_of_contents, read_section, find_related). Descriptions are clear and action-oriented (avg ~120 chars, within 10-1024 range). All parameters have type definitions and descriptions. Tool annotations (readOnlyHint, idempotentHint) are properly applied. However, output schemas are not documented in the source code, LLMs cannot see what fields to expect from responses. Parameter descriptions lack format constraints (e.g., path format, heading_id format). No error handling guidance visible. Missing pagination details for list_pages and search_docs despite accepting limit parameters.
Find documentation pages related to a given page based on shared tags.
Get the site-wide navigation tree showing all pages organized by directory.
List documentation pages with metadata. Optionally filter by tag.
Read a documentation page with metadata. Returns clean markdown with frontmatter stripped and metadata prepended.
Read a specific section of a page by heading anchor ID. Returns markdown from the heading to the next heading at the same or higher level.
Full-text search across all documentation pages. Returns ranked results with snippets. Supports tag: prefix syntax (e.g. "tag:deployment kubernetes") to filter by frontmatter tag.
Output schemas not documented. LLMs cannot see what fields responses contain, forcing them to guess at downstream tool parameters and breaking tool chaining.
Parameter descriptions lack format constraints. 'path' parameter accepts arbitrary strings up to 1024 chars but does not specify valid path format (e.g., 'relative paths only, no ../ traversal'). 'heading_id' format is undocumented.
Pagination not fully specified. search_docs and list_pages accept 'limit' but no cursor/offset or total_count guidance. Large result sets could exhaust context window.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | D | 59 | 2026-07-28+ | v2 |
No error handling guidance visible in tool definitions. If a path is invalid or a heading_id does not exist, LLMs have no recovery instructions.
search_docs tag: prefix syntax documented in description but not enforced via schema. LLMs may misuse the syntax or pass invalid tag filters.