Analyze SMS phishing with the AI action

From the submitter: Employees can use this simple form to upload a screenshot of a suspicious text message. We use Claude Sonnet multi-modal input to analyze the content, extract IOCs, and provide an initial triage to the employee. A case is created for the SOC, and escalated if multiple text messages are reported with matching numbers or URLs. We utilize prompt engineering from the paper "Principled Instructions Are All You Need", to improve AI output.