refund-radar

What This Skill Does

Refund Radar is a specialized financial analysis agent designed to audit your personal banking transactions and credit card statements. By leveraging advanced pattern recognition, the skill automatically scans CSV exports or raw transaction text to identify recurring subscription patterns, flagging suspicious activity, duplicates, or unexplained fees. It acts as an intelligent auditor, processing complex financial data into a clean, interactive HTML report that provides clear insights into your monthly spending habits.

Installation

To integrate this skill into your OpenClaw environment, execute the following command in your terminal: clawhub install openclaw/skills/skills/andreolf/refund-radar

Use Cases

• Subscription Audit: Quickly identify every recurring charge in your bank statement and determine if you are paying for unused services.
• Dispute Assistance: Automatically generate professional, clear, and concise templates for disputing unauthorized or incorrect transactions with your bank.
• Transaction Monitoring: Proactively manage your budget by identifying 'fee-creep' or sudden spikes in recurring costs.
• Financial Health Check: Get an overview of your spending patterns over a specific period, categorized by merchant, recurrence frequency, and severity of flagged issues.

Example Prompts

1. "Scan my bank statement for refunds: I uploaded the CSV, can you tell me if there are any duplicate charges from last month?"
2. "Audit my subscriptions and show me a report of all recurring charges that exceed $20."
3. "Help me dispute a charge: I see a $50 transaction from a merchant I don't recognize. Can you draft a refund request for me?"

Tips & Limitations

• Data Privacy: Ensure you strip sensitive PII (Personally Identifiable Information) like full account numbers or home addresses before sharing raw text files with the agent if you prefer maximum privacy.
• Parser Flexibility: The tool is highly adaptive to various formats, but standard CSV exports from major banks like Chase, Apple, and Mint yield the most accurate results.
• False Positives: Flagged items (like new merchant alerts) may occasionally trigger for legitimate one-time purchases; always review the HTML report manually before submitting formal disputes to your bank. The tool works best when provided with at least 30-60 days of transaction data to accurately baseline your recurring habits.