Back to skill
Skillv1.0.4
VirusTotal security
AgentPress · External malware reputation and Code Insight signals for this exact artifact hash.
Scanner verdict
SuspiciousApr 29, 2026, 4:36 AM
- Hash
- 93c3ab8437992f616fadb0ee5346af222e1f305937634eeaef9710d01f4643c3
- Source
- palm
- Verdict
- suspicious
- Code Insight
- Type: OpenClaw Skill Name: agentpress Version: 1.0.4 The skill is designed to execute an external `press` CLI binary, which introduces a supply chain risk if the `press` binary itself is compromised. More critically, the skill instructs the AI agent to pass user-provided input (e.g., post titles, descriptions) directly to `press` commands. This creates a shell injection vulnerability if the AI agent's execution environment does not adequately sanitize user input, potentially leading to arbitrary command execution. While the `SKILL.md` includes robust safety instructions (e.g., restricting file access, preventing secret exposure, multi-layer account deletion confirmation), the core mechanism involves risky capabilities (external command execution with user input) that classify it as suspicious due to the inherent RCE vulnerability risk.
- External report
- View on VirusTotal