PlayQuery defines 3 tools with basic schemas and descriptions. All tools have input schemas with type definitions and descriptions, meeting minimum requirements. However, descriptions are generic and lack LLM-optimized guidance on when/why to use each tool. The 'research' tool description (194 chars) is adequate but the other two are minimal (batch_search: 47 chars, batch_scrape: 47 chars). No output schemas documented. No error handling guidance. No parameter constraints (enums, ranges). Tools lack composition hints and chaining IDs. No tool annotations (readOnlyHint present via Risk field but not in MCP schema). Naming is clear (verb_noun pattern) but descriptions don't explain dependencies or prerequisites.
Fetch and extract content from multiple URLs.
Search the web for relevant URLs and snippets.
Research a topic and return a thorough, cited answer. Use this tool to get a well-researched answer to any question or topic. Provide a clear question or research request — not a search query. The tool will search the web, read the most relevant pages, and synthesise a response with inline numbered citations and a References section.
Minimal descriptions for batch_search and batch_scrape (47 chars each). Below 50-char baseline for discovery. LLMs cannot determine when to use batch_search vs research or batch_scrape vs research without richer context.
No output schemas documented. LLMs cannot plan downstream tool calls or extract structured data. batch_search and batch_scrape return types are opaque.
No error handling guidance. Tools do not explain what to do on failure (retry, ask user, fatal). No recovery hints like 'If search fails, try a simpler query.'
No parameter constraints. batch_search 'queries' array has no min/max length. batch_scrape 'urls' array has no length limit. Unbounded arrays invite LLMs to pass absurd values (1000 URLs, 500 queries).
Inferred effective spec: 2026-07-28+.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 44 | 2026-07-28+ | v2 |
No tool composition hints. research tool description does not explain it calls batch_search and batch_scrape internally, or when to call batch_* directly vs research. Agents cannot reason about tool relationships.