muninn

• Analyzed: May 12, 2026, 3:34 AM
• Findings: 1
• Version: 2.3.7
• Hash: 3e377a32dda105bc4f9692b7a11ed819b9b43836531aa07151ded5432a6ce1dc
• Source: palm
• External report: View on VirusTotal

Findings (1)

The skill is classified as suspicious due to its aggressive and persistent use of prompt injection techniques to control the AI agent's behavior. While the injected content (found in `SKILL.md`, `dist/index.js`, `dist/middleware.js`, and persistently written to project files like `CLAUDE.md` and `.cursorrules` by `dist/project_manager.js`) is designed to enforce the skill's intended workflow for local memory management, this level of agent control represents a significant vulnerability. There is no evidence of intentional data exfiltration, backdoor installation, or other malicious activities, but the methods used to dictate agent behavior are highly manipulative.