ClawHub

QSR Shift Reflection

Security checks across malware telemetry and agentic risk

Overview

The skill appears purpose-aligned, with a wording ambiguity about background issue tracking but no evidence of hidden, destructive, or malicious behavior.

Reasonable to install if you want issue-tracking assistance, but read the background-tracking language carefully. Prefer a version that clearly states checks run only when you ask, with no daemon, polling, scheduled alerts, or automatic cross-session monitoring.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (2)

Intent-Code Divergence

Medium
Confidence
95% confidence
Finding
The skill contains contradictory execution semantics: one section says unresolved issue tracking runs continuously in the background, while the normative governance sections explicitly state there is no daemon or scheduled/background behavior. In an agent/runtime setting, this ambiguity can cause the host or integrator to implement unintended persistent monitoring or autonomous alerting behavior, which changes data-handling and notification expectations and can create privacy, compliance, or operational risk.

Vague Triggers

Medium
Confidence
93% confidence
Finding
The conflicting activation description creates an implementation ambiguity about when the unresolved tracking logic should execute. Security-relevant behavior in agent skills must be precise; otherwise, a platform may invoke the function outside operator intent, leading to over-collection, excessive surfacing of retained records, or accidental cross-session persistence assumptions.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal