Nano Triple

• Outcome: Review
• Latest audit: May 27, 2026 at 9:17 PM UTC
• Version:
  1.2.0

Overview

The skill does what it advertises, but it tells the agent to pass raw user image prompts into shell commands without any escaping or safe argument-handling guidance.

Install only if you trust the external Nano Banana Pro helper it calls and your agent safely passes prompt text as arguments rather than interpolating it into a shell command. Expect each request or refinement to use three Gemini image generations and avoid putting secrets or sensitive data in prompts.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Trigger Abuse: Overly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access
• Supply Chain: Unpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (2)

Vague Triggers

Medium

• Confidence: 93% confidence
• Finding: The trigger phrases are very broad and likely to activate on many ordinary image-generation requests, causing this skill to run in situations where the user may not expect its specific behavior. Because the skill always generates three images in parallel, overbroad matching can lead to unnecessary API consumption, higher cost, and unintended invocation frequency.

Missing User Warnings

Low

• Confidence: 89% confidence
• Finding: The skill behavior automatically performs three parallel generations, but the user-facing description and flow do not clearly warn about the increased API usage or possible cost implications before execution. This creates a transparency and consent issue, especially in environments where image generation is metered or rate-limited.

VirusTotal

65/65 vendors flagged this skill as clean.