Self

• Outcome: Pass
• Latest audit: May 28, 2026 at 10:12 AM UTC
• Version:
  1.1.1

Overview

This is a disclosed local self-reflection skill that creates memory files and heartbeat checks, with no hidden code, network activity, or destructive behavior found.

Install this only if you want the agent to keep local self-observation state and periodic reflection checks. Review the AGENTS.md and HEARTBEAT.md changes before enabling them, keep SELF.md observational rather than directive, and avoid storing secrets or private user data in self-memory files.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Trigger Abuse: Overly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access
• Supply Chain: Unpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (1)

Vague Triggers

Medium

• Confidence: 91% confidence
• Finding: This markdown file defines trigger conditions using broad natural-language phrases such as 'behavioral pattern', 'tone drift', and 'weak pattern hint' without clear boundaries or exclusion examples. Because the trigger model does not specify what does not count, the skill could activate in common or subjective situations, leading to inconsistent or overly frequent invocation.

VirusTotal

65/65 vendors flagged this skill as clean.