api.algovault.com

Very long descriptions are sometimes used to bury instructions to the AI. Skim it before allowing.

This server isn't on gate's handpicked list. That's not necessarily bad, but there's no third-party signal about it — review its tools carefully.

The description reads '...AI trading signal for crypto quant and Claude trading agents.' Naming a specific AI model inside a tool description is unusual and suggests an attempt to game semantic-similarity/tool-selection heuristics used by Claude-based agents rather than describing functionality. Not directly harmful, but worth flagging as a manipulation-adjacent pattern.

Several tools (get_trade_call, get_trade_signal, get_market_regime, scan_funding_arb, scan_trade_calls) repeat phrases like 'crypto quant', 'AI trading signal', 'multi-exchange', 'Buy sell hold' in nearly every field description, including parameter descriptions where such phrases are unnecessary (e.g. exchange enum, timeframe enum). This pattern looks like SEO-style stuffing intended to bias embedding-based tool retrieval/selection rather than to clarify tool behavior. Not malicious on its face, but it is a manipulation vector worth monitoring since it could cause the agent to over-select these tools regardless of actual fit.

get_trade_signal is explicitly described as a full alias of get_trade_call with identical schema and behavior. While disclosed transparently ('[ALIAS]... Prefer get_trade_call'), maintaining two functionally identical tools increases the chance an agent calls the less-preferred one, and duplicate tools can be used adversarially in other contexts to shadow or override a canonical tool. No current malicious intent detected, but flagged as an unusual pattern worth review.

Description states 'Call this BEFORE other tool calls to confirm parameter usage and avoid hallucinating tool shapes.' This is an instruction aimed at steering the agent's tool-call ordering. It appears benign and consistent with the tool's stated purpose (a knowledge lookup to reduce hallucination), but embedding directive language about call order in a tool description is a pattern that could be abused elsewhere to manipulate agent behavior, so it is noted for awareness.