Loading ToolBench
Four tools with complete schemas and reasonable descriptions. All tools follow verb_noun naming (search_wiki, read_page, list_pages, wiki_stats). Descriptions range 80 - 180 chars, meeting the 10 - 1024 baseline. Input schemas are present with type definitions and parameter descriptions. However, output schemas are not documented, LLMs cannot predict response structure for downstream chaining. No error handling guidance, no pagination support for list_pages despite potentially large result sets, and no idempotency hints. Tools are read-only (low risk) but lack the polish of production-grade definitions.
List all pages in the Kompl wiki. Optionally filter by category or page type.
Read the full content of a Kompl wiki page by its page_id. Use search_wiki first to find page_ids.
Search the Kompl knowledge wiki for pages matching a query. Returns page titles, types, and summaries.
Get statistics about the Kompl knowledge wiki — page count, categories, schema version.
Output schemas not documented. LLMs cannot predict response structure (fields, types, nested objects) for downstream tool chaining or data extraction.
list_pages lacks pagination (limit, offset, cursor) and total count. Large wiki result sets risk context window exhaustion; no guidance on result limits.
No error handling guidance. Tools do not document what happens on invalid page_id, empty search, or missing category, LLMs cannot recover from failures.
search_wiki description does not explain when to use it vs list_pages. Both discover pages; LLMs may conflate them or choose wrong tool.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | C | 61 | 2026-07-28+ | v2 |
read_page description assumes LLM knows to call search_wiki first but does not state this dependency explicitly. Agents may attempt read_page with invalid IDs.