Back to skill
Skillv0.6.3
VirusTotal security
ClawdCursor · External malware reputation and Code Insight signals for this exact artifact hash.
Scanner verdict
BenignApr 30, 2026, 4:21 AM
- Hash
- 2b961be0a04cb570f1ac7eb6d845252e0b8aa842a21a68bd171594b31d9a2995
- Source
- palm
- Verdict
- benign
- Code Insight
- Type: OpenClaw Skill Name: clawd-cursor Version: 0.6.3 The OpenClaw AgentSkills skill 'clawd-cursor' is designed for desktop GUI automation, providing the AI agent with capabilities to interact with applications, type, click, and navigate. While these capabilities are inherently powerful and could be misused, the `SKILL.md` documentation explicitly and repeatedly instructs the AI agent on critical safety measures, user consent, and privacy. It mandates asking the user for approval before sensitive actions (e.g., email, banking, deleting), binds its API to `127.0.0.1` only, and states that data sent to cloud AI providers goes only to the user's chosen provider, not to skill authors. There is no evidence of intentional data exfiltration to unauthorized parties, persistence mechanisms, or obfuscation, and the prompt injection surface is used to enforce safety rather than bypass it.
- External report
- View on VirusTotal