Dlazy Gpt 5.5
Security checks across malware telemetry and agentic risk
Overview
The reviewed skill artifacts are coherent with repo maintenance, Convex development, UI proof, and moderation workflows, with powerful actions disclosed and mostly gated by user intent.
Install only if you want these repo-maintenance and Convex workflow helpers. Before using the autoreview helper, consider running it with --no-yolo if you do not want nested review to have full local access. For moderation workflows, verify you are using the intended ClawHub account and target because the actions can affect public content and user accounts.
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
62/62 vendors flagged this skill as clean.