Agent Architect

• Outcome: Pass
• Latest audit: May 29, 2026 at 8:25 PM UTC
• Version:
  1.0.0

Overview

This is an instruction-only agent-audit skill that recommends where changes should be made, with no code, credentials, install step, or hidden execution behavior.

Safe to install as an audit aid. Before using it, decide whether daily memory logging is acceptable, and replace the named 'Don approval' rule if it does not match your own governance process.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Rogue Agent: Self-Modification, Session Persistence
• 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)

Self-Modification

High

• Category: Rogue Agent

Content

If the fix changes **how the agent sounds**, patch persona.
If it changes **how the agent decides**, patch rules.
If it changes **what the agent retains or recalls**, patch memory.
If it changes **how the agent repeatedly executes a task category**, patch skills.

• Confidence: 85% confidence
• Finding: patch skill

VirusTotal

63/63 vendors flagged this skill as clean.