ADI Decision Engine

• Analyzed: May 11, 2026, 5:03 AM
• Findings: 1
• Version: 0.1.0
• Hash: 408f6b58a07825171eb79d8b672dcf83f0eb57f44c615180d9e342eab0f1bba3
• Source: palm
• External report: View on VirusTotal

Findings (1)

The OpenClaw AgentSkills skill bundle for the ADI Decision Engine is classified as benign. The `SKILL.md` provides clear instructions to the AI agent, including explicit 'Safety and honesty rules' that prohibit requesting API keys or requiring network access for the core workflow, mitigating common prompt injection risks. The Python scripts (`scripts/_runtime.py`, `scripts/normalize_problem.py`, `scripts/run_adi.py`, `scripts/validate_request.py`) securely execute external commands using `subprocess.run` with a list of arguments, preventing shell injection vulnerabilities. While the skill relies on an external `adi` CLI or Python package, which introduces a dependency-related supply chain risk, the skill's own code does not exhibit any malicious intent, data exfiltration, persistence mechanisms, or obfuscation.