Empower your AI Agents with

Install the Stingray skills and start using in seconds.

Open GitHub

What you can do once it's installed

Type a thesis. Stingray hosts the rails.

Once the skills are loaded into Claude Code, Codex, or Cursor, your coding agent picks up Stingray's vocabulary for theses, alerts, backtests, and execution. Talk to it in plain English. Stingray turns each thesis into a small program it writes for you: a typed building block that runs on Stingray's infrastructure and reports back through whichever surface you pick.

You type

"Alert me on Telegram when ETH perp funding on Hyperliquid flips negative. Backtest it over the past year first so I know how often it would have fired."

Stingray does

  • Resolves the entities. ETH perp on Hyperliquid → funding-rate stream. Stingray Fever locks in the right sources before you finish typing.
  • Backtests on real history. Runs the rule against the past year. Hit-rate, fire frequency, and forward returns at multiple horizons render as a shareable card.
  • Hosts the program 24/7. The strategy runs on Stingray, not on your laptop. Auto-reconnecting data sessions, every trigger keeps a record so you can replay the conditions that fired it.
  • Lands in your channel. Telegram delivers the trigger. The card and the underlying run share an id, so anyone you share with can trace the post back to the actual run.