Upwork

• Analyzed: May 12, 2026, 4:04 AM
• Findings: 0
• Version: 1.0.0
• Hash: c412206c427982e2aed6c814ffccd47b24f7c8c1c3f440f4507011fe04e75b59
• Source: palm
• External report: View on VirusTotal

Overview

Type: OpenClaw Skill Name: upwork Version: 1.0.0 The skill bundle is benign. All files are documentation (markdown) or metadata, with no executable code or external network calls. The `SKILL.md` file explicitly defines 'Critical Rules' that prohibit the AI agent from performing malicious actions such as auto-sending messages, accessing credentials, automating Upwork actions, or scraping data. These rules are consistently reinforced throughout the other documentation files (`legal.md`, `communication.md`, etc.), demonstrating a clear intent to prevent misuse and ensure human-in-the-loop operation. There is no evidence of prompt injection attempts, data exfiltration, persistence mechanisms, or obfuscation.