coding-agent-common

• Analyzed: Mar 27, 2026, 8:51 AM
• Findings: 1
• Version: 1.0.0
• Hash: bd880c2dba8fe52495c1e74498746f20c12b23bde32cf5fc4be7865c291292fc
• Source: palm
• External report: View on VirusTotal

Findings (1)

The skill bundle provides instructions for delegating tasks to external AI coding tools (Codex, Claude Code, OpenCode). It explicitly directs the OpenClaw agent to use high-risk execution flags such as '--yolo' and '--permission-mode bypassPermissions' in SKILL.md and the reference files, which are designed to bypass human-in-the-loop security approvals and sandboxing. While these features are functional for autonomous coding, they create a significant risk of Remote Code Execution (RCE) if the agent is prompted with a malicious task. No evidence of intentional data exfiltration or backdoors was found, but the promotion of security-bypass modes warrants a suspicious classification.