ClawHub
Back to skill

Security audit

Gemini Deep Research

Security checks across malware telemetry and agentic risk

Overview

This skill appears to do what it says: send user-requested research jobs to Gemini and save the resulting reports locally.

Install only if you are comfortable sending research prompts and any selected Gemini file-search context to Google's Gemini service. Use an environment variable rather than passing the API key on the command line on shared machines, monitor quota or cost for long-running jobs, and choose a private output directory when reports or response metadata may contain sensitive material.

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
83% confidence
Finding
The script persists both the generated report and the full API response to local disk by default, which may include sensitive user queries, proprietary research content, file-search-derived data, or metadata. Because this happens automatically without an explicit warning or opt-in, users may unintentionally leave sensitive material in shared directories, backups, or CI workspaces.

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.