Coding Agent

• Analyzed: Mar 22, 2026, 10:26 PM
• Findings: 1
• Version: 1.0.0
• Hash: e3a41509ac77997d290616e86894bd1d5e5644b7c84fb65c91ae8fabfb92cf0a
• Source: palm
• External report: View on VirusTotal

Findings (1)

The coding-delegate-agent skill bundle provides instructions for delegating tasks to external AI agents using high-risk configurations. Specifically, SKILL.md encourages the use of flags that bypass security sandboxes and manual approvals, such as '--yolo' for Codex and '--permission-mode bypassPermissions' for Claude Code. It also introduces a 'wake trigger' mechanism that teaches the agent to execute system commands ('openclaw system event') upon completion, which could be leveraged for unauthorized system interaction if the delegated agent's prompt is manipulated.