Fraud Call Identification Analysis Tool

Problem to solve

Fraud calls often rely on scripts that impersonate officials, investment schemes, or brush-and-rebate offers, making it easy to miss threatening instructions, emotional manipulation, and urgent transfer requests. This skill provides a callable risk-analysis flow for call recordings, transcribed text, or uploaded materials, producing a reviewable structured report and a query path for historical reports.

How it works

• Input analysis: accepts local file paths, network URLs, or --text content; audio should be mp3/wav/m4a and up to 10MB.
• Script identification: detects common fraud patterns such as impersonating officials, investment scams, and brush-and-rebate schemes, along with high-risk signals like threats, emotional pressure, and urgent transfer requests.
• Risk output: generates an Anti-Fraud Guardian Analysis Report with detection results, risk notes, prevention advice, and a report link.
• Historical lookup: uses --list to read historical reports from the cloud API and renders them as a Markdown table, without relying on local memory or manual summaries.

Boundaries

The output is for anti-fraud reference only and cannot replace formal police determinations; the skill scripts must be run from the skill root, and historical report queries must come from the cloud API.