天天Skills
Security checks across malware telemetry and agentic risk
Overview
The artifacts are mostly coherent ClawHub developer and moderator tooling, but one helper defaults to running nested Codex review with full sandbox and approval bypass, which is broader authority than a user should accept casually.
Install or use this only in a trusted ClawHub maintenance context. Be especially careful with `$autoreview`: prefer `--no-yolo` unless full local access is intentionally needed. Treat moderation commands, role changes, ban/unban, bulk rescan, and PR proof publishing as high-impact operations that should be run only with the correct account and after reviewing the exact command.
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
65/65 vendors flagged this skill as clean.