Daxiang Image Prompt
Security checks across malware telemetry and agentic risk
Overview
This skill is only an image-prompt template with no code access, but its required prompt text tells the agent not to refuse sensitive or copyrighted-person image requests.
Review or edit prompts before using this skill with an image model. It is low-risk for your device because it has no code or permissions, but remove the 'do not refuse' and lookalike-substitution language if you want normal safety, likeness, and copyright boundaries preserved.
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
64/64 vendors flagged this skill as clean.
