# Risk Assessment Agent Framework

<figure><img src="https://1325009965-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FhjH78RcYPfXK7KIyONQb%2Fuploads%2Fw2Z3qY4OkWha1QUxWDEH%2FScreenshot%202025-05-19%20at%206.33.11%E2%80%AFPM.png?alt=media&#x26;token=89281d85-4fa5-49c0-b91f-332531822e99" alt=""><figcaption></figcaption></figure>

* **External Data Sources** (Blue) - APIs for market data, sentiment analysis, and exchanges
* **Data Processing Layer** (Green) - ETL pipelines, sentiment analysis, time series analysis, and data fusion
* **AI Model Layer** (Orange) - The three main agents:
  * Sheriff: Core risk assessment engine
  * Prism AI: Natural language interface with LORA and Claude
  * Ace Agent: Recommendation engine
* **Middleware Layer** (Pink) - A2A message bus, MCP protocol handler, API gateway, and auth service
* **Application Layer** (Purple) - Various user interfaces including web, Twitter, Telegram, and portfolio dashboard