Security audit
Idea Trace
Security checks across malware telemetry and agentic risk
Overview
The artifact appears to be a coherent ClawHub development and maintenance workspace with disclosed local skills and no evidence of hidden or malicious behavior.
Install or use this only in a trusted ClawHub development context, with least-privilege GitHub/Convex/OpenAI credentials and care around staff moderation commands, since those can affect real registry users and content.
SkillSpector
By NVIDIA
Vulnerability Patterns
- Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
- Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
- Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
- Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
- Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
VirusTotal
VirusTotal findings are pending for this skill version.
Static analysis
No suspicious patterns detected.