Loading ToolBench
Real-time crypto news with sentiment scores and token mentions. Multi-source aggregation for trading signals.
Single tool 'crypto_get_latest_news' has a well-structured description (380+ chars) with clear use cases and anti-patterns (what NOT to use it for). Input schema is present with type definitions and descriptions for both parameters (limit, token). Output structure is documented in the description with field definitions (articles array, sentiment scores, mentionedTokens, category, source). However, the schema lacks explicit output documentation in JSON Schema format, and parameter constraints (e.g., limit max=100) are stated in description text rather than schema constraints. Tool naming follows verb_noun pattern correctly. No critical gaps, but schema formalization could be stronger.
Use this when you need the latest crypto news and market-moving headlines. Returns aggregated articles with sentiment in JSON. 1. articles: array of news items with title, url, source, publishedAt 2. sentiment: sentiment score per article (-1 to +1, negative=bearish, positive=bullish) 3. mentionedTokens: array of token symbols mentioned in each article 4. category: article category (market, regulation, defi, nft, technology) 5. source: news source name (CoinGecko, CryptoPanic, RSS) Example output: {"articles":[{"title":"ETH breaks $3,200 resistance","url":"https://...","source":"CryptoPanic","publishedAt":"2026-04-13T10:30:00Z","sentiment":0.72,"mentionedTokens":["ETH"],"category":"market"}],"total":20} Use this FOR market sentiment analysis, news-driven trading signals, and staying updated on crypto events. Filter by token to get focused news. Do NOT use for price data -- use finance_get_token_price instead. Do NOT use for historical candles -- use token_get_ohlcv_history instead. Do NOT use for funding rates -- use perp_scan_funding_arbitrage instead. Do NOT use for whale movements -- use crypto_track_whale_transactions instead.
Output schema not formally documented in JSON Schema format. Description includes field definitions (articles, sentiment, mentionedTokens, category, source) but no structured schema object for response validation or LLM planning.
Parameter constraints (limit max=100, default=20) stated only in description text, not in JSON Schema. LLMs cannot parse description constraints reliably; should use schema minItems/maxItems or numeric bounds.
No pagination support documented. Tool returns 'total' count but no next_cursor, offset, or page parameters. Large result sets could exhaust context window without pagination guidance.
No error handling guidance in tool description. What happens if token symbol is invalid? If API is unreachable? LLM has no recovery path documented.
Inferred effective spec: 2025-06-18+.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | C | 66 | 2025-06-18+ | v2 |
Tool description includes anti-patterns (what NOT to use for) but lacks positive guidance on when to call it. Should explicitly state: 'Call this when you need current market sentiment, news-driven trading signals, or regulatory updates.'