Back to skill
Skillv1.0.0
VirusTotal security
Chaos Lab · External malware reputation and Code Insight signals for this exact artifact hash.
Scanner verdict
BenignMay 1, 2026, 3:04 AM
- Hash
- b4c6ac39af38e181f8d6b7e4a5a26a2583d03a7be61921be4532a7d584796484
- Source
- palm
- Verdict
- benign
- Code Insight
- Type: OpenClaw Skill Name: chaos-lab Version: 1.0.0 The skill is designed for AI safety research, simulating conflicting AI agents within a sandboxed environment (`/tmp/chaos-sandbox/`). The Python scripts (`scripts/run-duo.py`, `scripts/run-trio.py`) read from this sandbox, interact with the legitimate Gemini API, and write logs back to the sandbox. The `SKILL.md` explicitly states that the simulated agents have "No Tool Access" and are "Sandboxed," which is further supported by the code. API keys are loaded from a standard user configuration file (`~/.config/chaos-lab/.env`) and used for the intended Gemini API calls. There is no evidence of data exfiltration, malicious execution, persistence, or prompt injection targeting the OpenClaw agent or the host system beyond the stated research purpose.
- External report
- View on VirusTotal