input classification

Security checks across malware telemetry and agentic risk

Overview

This is a documentation-only skill for classifying and routing user requests, with no evidence of hidden execution or data exfiltration.

Safe to install as a classification and routing aid. Before using it with sensitive requests, confirm how your agent platform implements its logging, human-review queues, encryption, access controls, and retention policy.

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

Medium

Confidence: 85% confidence
Finding: This markdown file falls under SQP-1, and the section title 'State Transition Trigger' plus the statement 'After classification, set the appropriate state transition' describes a trigger-like mechanism without clearly scoping when it should or should not activate beyond the broad condition 'after classification.' The document does not provide exclusion conditions or negative examples, which could make downstream invocation logic overly broad.

VirusTotal

63/63 vendors flagged this skill as clean.

View on VirusTotal