Integration with AI signals

AI Agent Webhook Integration (via Pre-Built Prompts)

Overview Implement native webhook support allowing users to connect Large Language Models (LLMs) and custom AI agents (e.g., Claude, custom GPTs) directly to the SharkSignals execution engine. The feature relies on providing users with a pre-built system prompt that dictates the required JSON schema, bypassing the need for users to write custom integration code.

Core Functionality & Acceptance Criteria

• Optimized Webhook Endpoints: Ensure the webhook ingestion pipeline cleanly accepts and parses LLM-generated JSON payloads, with error handling for common LLM syntax hallucinations.

• Pre-Built Prompt Template: Create and maintain a standardized system prompt template that users can copy/paste into their preferred AI agent. The prompt must instruct the model on exactly how to structure, format, and transmit trade requests.

• Authentication: All AI-generated requests must be secured using the user's existing 64-character secret key.

• Execution Logic: Payloads received from AI agents must seamlessly integrate with existing architectural rules, including single-entry order generation and linked multi-bracket routing (TP/SL).

Primary Use Case Allows users to execute trades based on natural language processing, sentiment analysis, or macroeconomic data parsed by an AI, using a simple "plug-and-play" prompt approach rather than custom API development.