Adaptive Reasoning

Security checks across malware telemetry and agentic risk

Overview

This instruction-only skill broadly influences reasoning behavior, but its behavior is disclosed, purpose-aligned, and does not request code execution, credentials, file access, or data transfer.

Install this only if you want an always-on assistant behavior modifier that can silently choose deeper reasoning and add visible reasoning icons. Avoid it for workflows that require strict output formatting, minimal token use, or explicit approval before session behavior changes.

SkillSpector

By NVIDIA

Vulnerability Patterns

Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
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

Findings (1)

Vague Triggers

High

Confidence: 92% confidence
Finding: Triggering on every user message gives this skill blanket influence over all interactions, including low-risk or unrelated prompts. Because it attempts to silently alter reasoning/session behavior and append output markers automatically, broad activation increases the chance of unintended behavior, policy interference, prompt-surface expansion, and instruction conflicts across the entire session.

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal