Loading ToolBench
A collection of agent skill implementations including MCP tools for various domains (stock research, database design, brainstorming, web testing, etc.)
Four tools with basic verb-noun naming (get_*) and minimal descriptions (15-60 chars). All tools have input schemas with typed parameters and descriptions, but lack output schema documentation. No error handling guidance, no parameter constraints (enums), and no indication of idempotency or side effects. Tools operate on SEC financial data (read-only), but descriptions do not state this explicitly. Missing dependency hints (e.g., 'Call get_cik first if you only have a ticker'). No tool annotations (readOnlyHint, etc.). Schemas are present but sparse, no examples, no pagination support, no result limits documented.
Resolve ticker to zero-padded SEC CIK
Fetch raw EDGAR companyfacts payload for a ticker
Return recent SEC filing metadata for common report forms
Extract and normalize financial statement data from company facts
Output schemas not documented. LLMs cannot plan downstream calls or extract required fields without knowing what each tool returns.
Descriptions are too brief (15 - 60 chars) and lack context. Do not state WHEN to use each tool, what it returns, or how it differs from similar tools. Descriptions should be 50 - 200 chars with actionable guidance.
No parameter constraints (enums, ranges, patterns). 'ticker' accepts any string; no validation that it is a valid stock symbol. LLMs will pass invalid tickers, causing API errors.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | D | 55 | 2026-07-28+ | v2 |
No error handling guidance. If a ticker is not found or API fails, LLMs have no recovery path. Errors should guide the agent: 'Ticker not found. Try search_tickers() or verify the symbol.'
No tool annotations (readOnlyHint, destructiveHint, idempotentHint). All four tools are read-only and idempotent, but this is not declared. Agents cannot infer safety properties.