Back to skill
Skillv1.4.8
VirusTotal security
HardStop · External malware reputation and Code Insight signals for this exact artifact hash.
Scanner verdict
BenignApr 30, 2026, 4:23 AM
- Hash
- e923a9b2f5247929c425fd026d48cc606e0320fec5713590d2f4cb18b36b1a69
- Source
- palm
- Verdict
- benign
- Code Insight
- Type: OpenClaw Skill Name: hs Version: 1.4.8 The OpenClaw AgentSkills skill bundle is designed to act as a 'Hardstop' safety mechanism, instructing the AI agent to detect and block dangerous shell commands, file reads, and potential prompt injection attempts. The `SKILL.md` explicitly lists common malicious patterns (e.g., reverse shells, data exfiltration, system destruction, encoded payloads) and instructs the agent to block them or seek explicit user confirmation for risky actions. The only external execution instructed is `python ~/.claude/plugins/hs/commands/hs_cmd.py`, which is explicitly stated to be a local script of an installed plugin, not remote code. There is no evidence of intentional harmful behavior; rather, the skill's purpose is to prevent such actions.
- External report
- View on VirusTotal