Loading ToolBench
A tech news aggregator server that fetches the latest news articles from specified tech news sites
Single tool with significant definition gaps. Tool naming is acceptable (verb_noun pattern) but description is generic and lacks LLM-optimized guidance. Input schema exists but is minimal with no enum constraint despite documentation listing specific allowed values. Output is unstructured string rather than typed response. No error handling guidance. No pagination support. Lacks parameter descriptions in schema (only in docstring). No output schema documentation. Does not meet production baselines for A/B/C grade tools.
Fetches the latest news articles from a specified tech news site. Args: site (str): The tech news site to fetch articles from. Options are 'techcrunch', 'theverge', 'wired', 'engadget', 'arstechnica'. Returns: str: A summary latest news articles with their titles and summaries.
Input parameter 'source' lacks enum constraint despite tool docstring listing exactly 5 allowed values (techcrunch, theverge, wired, engadget, arstechnica). Free-form string invites hallucinated values and reduces clarity.
Output is unstructured string ('A summary latest news articles with their titles and summaries') but actual implementation returns raw text blob from web scraping. No structured schema (e.g., list of {title, summary, url, published_date}). LLMs cannot reliably parse or chain results.
Error handling returns generic 'Timeout error' string on all exceptions. Does not follow recovery-guide pattern, LLM has no guidance on retry eligibility, root cause, or next steps.
Tool description (194 chars in docstring) is generic and does not explain WHEN to use it vs alternatives, what structure is returned, or dependencies. Docstring is longer than description field itself, indicating schema is incomplete.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-22 | F | 40 | 2026-07-28+ | v2 |
| 2026-03-09 | F | 35 | - | v1 |
No pagination support. Tool silently truncates to first 5 paragraphs and joins into single string. If result is large, context window overhead is not managed. No limit parameter, no offset/page parameter, no total count returned.
Parameter description in schema is identical to docstring arg description ('The tech news site to fetch articles from...') but schema 'source' param key does not match docstring 'site' param name. Mismatch creates confusion about actual parameter naming.
No result limit enforcement or documentation. Tool scrapes and returns whatever BeautifulSoup extracts; no cap on token cost or context window impact. Production baseline is 20-50 items max with pagination.