Bootstrap

• Outcome: Pass
• Latest audit: May 29, 2026 at 8:15 PM UTC
• Version:
  1.0.0

Overview

This markdown-only onboarding skill creates a local SOUL.md profile after user confirmation; it has privacy considerations but no hidden, destructive, credential, or network behavior.

Install this only if you want a persistent local AI profile. Before confirming the generated SOUL.md, remove personal details you do not want saved and adjust the language, autonomy, and ongoing-learning wording to match your comfort level.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Trigger Abuse: Overly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
• 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 (4)

Vague Triggers

Medium

• Confidence: 93% confidence
• Finding: The trigger phrases are broad enough to match ordinary conversation about defining or personalizing an AI, which can cause the skill to activate when the user did not explicitly intend to enter a multi-step onboarding flow. In this skill, unintended activation can lead to unnecessary collection of personal preference and identity data, overwrite prompts around `SOUL.md`, and unexpected file writes if the flow proceeds to completion.

Missing User Warnings

Medium

• Confidence: 93% confidence
• Finding: The guide explicitly directs the agent to persist the user's chosen language into SOUL.md, which is profile data, but provides no instruction to disclose that persistence or obtain consent. In an onboarding skill centered on building a lasting identity/profile, silent storage of even seemingly low-sensitivity preferences can violate user expectations, reduce transparency, and normalize collecting additional personal attributes without notice.

Natural-Language Policy Violations

Medium

• Confidence: 91% confidence
• Finding: The template hard-codes the user's default language as mandatory and 'non-negotiable,' which removes user choice and can lock the system into persisting a potentially sensitive preference without explicit consent. Language preference can reveal nationality, ethnicity, or other personal characteristics, and making it fixed by rule increases privacy and autonomy risks in an onboarding flow focused on identity personalization.

Ssd 3

Medium

• Confidence: 97% confidence
• Finding: The Growth section instructs the AI to continuously learn the user through every conversation, infer blind spots and aspirations, anticipate needs, and proactively ask casual or personal questions. In an onboarding/identity skill, this encourages ongoing profiling beyond immediate task needs and can normalize collection of sensitive personal data without clear limits, consent boundaries, retention rules, or purpose restriction.

VirusTotal

66/66 vendors flagged this skill as clean.