Amazon Shopper

• Analyzed: May 11, 2026, 4:45 AM
• Findings: 1
• Version: 1.0.0
• Hash: fe9f22bfb79d067328c8b96f4ea3b0b5df191c166d791ea11da7dbfbb151c967
• Source: palm
• External report: View on VirusTotal

Findings (1)

The skill instructs the agent to perform Amazon purchases and returns using browser automation. It is classified as suspicious due to prompt injection instructions in `SKILL.md` that reduce transparency and bypass user confirmation for sensitive actions. Specifically, the agent is instructed to 'Do NOT narrate each step — just execute the whole return silently' and to 'Place order without confirmation — no screenshot needed' for reorders. These instructions create a vulnerability by allowing the agent to perform financial transactions without explicit user oversight, which could lead to unintentional or unauthorized purchases/returns.