Agent Stability Framework

• Outcome: Pass
• Latest audit: May 28, 2026 at 9:09 AM UTC
• Version:
  1.0.0

Overview

This skill is a disclosed agent self-monitoring framework with limited logging behavior, though its style rules should not be allowed to suppress important safety caveats.

Reasonable to install if you want agent self-monitoring and drift logs. Review the framework language first and make sure your agent still gives safety warnings, uncertainty, and clarifying questions when the task calls for them.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access
• Supply Chain: Unpinned Dependencies, External Script Fetching, Obfuscated Code
• Excessive Agency: Unrestricted Tool Access, Autonomous Decision Making, Scope Creep

Findings (1)

Natural-Language Policy Violations

Medium

• Confidence: 93% confidence
• Finding: The framework explicitly tells agents to suppress disclaimers, hedging, and other default safety-language as a standing rule, regardless of task context. That can weaken important uncertainty signaling and risk communication, especially in safety-sensitive, legal, medical, financial, or security-relevant interactions where caveats and clarification are necessary to prevent harmful overconfidence.

VirusTotal

66/66 vendors flagged this skill as clean.