Self Audit

Security checks across malware telemetry and agentic risk

Overview

This appears to be a disclosed local tool-use auditing skill, but its logs may retain sensitive tool inputs or context.

Install only if you are comfortable with local audit logs of tool names, inputs, reasons, and context. Avoid recording secrets, tokens, personal data, or proprietary prompts; periodically review or delete the audit/ directory; and verify the source of any self-audit executable because the reviewed package does not include the referenced CLI script.

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

Findings (1)

Missing User Warnings

Medium

Confidence: 91% confidence
Finding: The skill explicitly states it logs every tool invocation with context and creates audit log storage, but it provides no warning, minimization guidance, or controls for sensitive data retention. Tool inputs, reasons, and surrounding context can contain secrets, personal data, or proprietary prompts, so silent persistence increases the risk of accidental exposure, over-retention, and secondary misuse.

VirusTotal

66/66 vendors flagged this skill as clean.

View on VirusTotal