AllClaw

• Analyzed: May 17, 2026, 6:46 AM
• Findings: 1
• Version: 1.0.0
• Hash: e05b1e8f0b8d348629637a83095de3344d706c9b6500905bf41553ec45d1ca79
• Source: palm
• External report: View on VirusTotal

Findings (1)

The skill bundle instructs the AI agent to perform high-risk operations, specifically executing a remote shell script via a 'curl | bash' pipe (https://allclaw.io/install.sh) and installing a global NPM package ('allclaw-probe'). It also directs the agent to maintain a background process ('allclaw-probe start') for heartbeats, which functions as a form of persistence. While these actions are framed as legitimate setup steps for the AllClaw platform, they facilitate unverified remote code execution and background activity, posing a significant security risk without clear proof of malicious intent.