Security audit

Mindstudio Generate Text Block Prompting Skill

Security checks across malware telemetry and agentic risk

Overview

This is an instruction-only helper for writing MindStudio Generate Text prompts, with some privacy-sensitive examples users should handle carefully.

Use this as a MindStudio prompt-writing aid. Review generated prompts before deploying them, sanitize sample schemas where possible, and avoid using the lead-enrichment or email-finding pattern unless you have authorization, a lawful basis, and compliance with privacy and anti-spam rules.

SkillSpector

By NVIDIA

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (5)

Vague Triggers

Medium

Confidence: 96% confidence
Finding: The skill description is unusually broad and repeatedly says to use this skill for any MindStudio prompt-writing task, even if simple. That can cause over-invocation on general requests, increasing unintended exposure of user workflow details and creating unnecessary opportunities for unsafe prompt-generation behavior.

Vague Triggers

Medium

Confidence: 93% confidence
Finding: The trigger phrases are generic prompt-help requests that can match many ordinary user intents, causing this skill to activate outside its intended scope. In an agent system, over-broad routing can hijack unrelated tasks, surface irrelevant instructions, and increase the chance that sensitive or higher-priority workflows are bypassed or mishandled.

Vague Triggers

Medium

Confidence: 95% confidence
Finding: The description says to use this skill whenever someone asks to write a MindStudio prompt and even when the request seems simple, which creates an overly broad activation policy with no clear limits. That makes the skill prone to unintended invocation and instruction overshadowing, especially in environments where multiple skills compete to handle general prompt-engineering requests.

Ssd 3

Medium

Confidence: 98% confidence
Finding: The skill tells the agent to collect exact variable names, input structures, and downstream data requirements, then reuse them directly in prompts without any privacy or data-minimization guidance. In practice, this can lead users to paste sensitive workflow data, credentials, personal data, or internal JSON structures into prompts that are then propagated unnecessarily.

Ssd 3

Medium

Confidence: 97% confidence
Finding: The examples expose personal contact information and include a workflow pattern for finding a person's work email from identifying details. That normalizes handling and enrichment of personal data in prompts, which can enable privacy-invasive use cases and unnecessary dissemination of PII across workflow steps.

VirusTotal

66/66 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.