santiment

Execute trading signals and monitor positions, orders, and balances on the Hyperliquid DEX through the `hyperhandler` CLI. Use when an agent or user needs to place/cancel orders, read account state, or manage HD wallets/keys for Hyperliquid from a shell with machine-readable JSON output.

Query and mutate Santiment Score data (predictions, markets, portfolio, leaderboards, issuers, stakes) through the `score` CLI. Use when an agent or user needs to read or write Score product data from a shell with machine-readable JSON output.

<verb + object>. Use when <trigger phrase>.

Fetch Santiment crypto metrics with stable defaults. Prefer san.get and san.get_many. Use GraphQL only for metadata inspection, custom selectors, or custom response shapes.

Read how the crowd is POSITIONED around a coin during a social spike and whether that positioning is EXTREME vs the coin's own history — not a news summary. Use when a token is trending, its social volume / dominance is spiking (now or at some past date), or someone asks "what happened with X / why is it being talked about / is this hype real / is the crowd long or short here / who is driving it". Turns raw `social_messages` into four signals: (1) extremeness vs a trailing baseline (percentiles, not adjectives), (2) organic vs manufactured (copies, channel/author concentration, acceleration shape), (3) narrative-vs-chain divergence (the crowd's concrete claims tested against on-chain flows), (4) crowd price-level clusters (the support / targets people actually name). Surfaces the dominant trending topics driving the spike (the "what happened") and delivers a positioning VERDICT with denominators and source links, then answers follow-up questions on the same pulled data.