Agent Sequential Thinking

• Outcome: Pass
• Latest audit: May 29, 2026 at 8:32 PM UTC
• Version:
  1.0.1

Overview

This is an instruction-only planning skill for breaking down complex tasks, with no executable code or credential access, but users should control when it runs and whether reasoning notes are saved.

Install it as a reasoning helper, not as an always-on behavior change. Prefer explicit invocation for sensitive work, and avoid saving local thinking logs for secrets, private data, confidential business details, or regulated legal/medical/financial matters unless you know where the files are stored and how to remove them.

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: 90% confidence
• Finding: The auto-trigger conditions are broad and subjective, such as tasks with multiple steps, complexity over an unspecified threshold, or users saying 'think about it.' In an agent system, this can cause unintended invocation of the skill, leading to unnecessary persistence of reasoning artifacts, expanded tool use, or altered agent behavior without clear user consent.

VirusTotal

64/64 vendors flagged this skill as clean.