AI Usage Audit — Monthly Retrospective & Insights

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

Overview

This skill openly analyzes recent AI chat history to create a usage report, but users should treat the resulting report as sensitive.

Install only if you are comfortable letting the agent review recent AI conversations. Use a narrower date range or project scope when possible, avoid running it over confidential chats, and review the saved HTML report before sharing or retaining it.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Trigger Abuse: Overly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
• 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

Findings (5)

Missing User Warnings

Medium

• Confidence: 92% confidence
• Finding: The prompt explicitly instructs the AI to pull 'as many as possible' conversations from the past month and analyze them, but the README provides only a functional requirement note and does not clearly warn users about the privacy sensitivity of aggregating chat history. Centralizing large volumes of prior conversations can expose sensitive personal, business, credential-adjacent, or regulated data to broader processing than the user may expect.

Vague Triggers

High

• Confidence: 96% confidence
• Finding: The trigger conditions are excessively broad and include casual phrases like "what have I been doing lately" and "let's do a retro," which can match ordinary conversation without clear consent to audit prior chats. In this skill, unintended activation is especially sensitive because execution would pull conversation history and analyze it, creating a privacy-impacting action from ambiguous user intent.

Vague Triggers

Medium

• Confidence: 90% confidence
• Finding: The skill says it should trigger when the user wants to understand usage efficiency or optimize collaboration, but it does not define clear boundaries for when this intent is present versus when the user is making a general reflective statement. That ambiguity increases the chance of the skill inferring consent and starting history analysis when the user did not explicitly request retrospective access.

Missing User Warnings

Medium

• Confidence: 95% confidence
• Finding: The skill instructs the agent to pull recent conversation history and optionally search past conversations, but it does not require a prominent user-facing notice or consent checkpoint before accessing that data. Because chat history can contain sensitive personal, business, or project information, silent retrieval materially increases privacy risk.

Missing User Warnings

Medium

• Confidence: 94% confidence
• Finding: The skill requires generating and saving a polished HTML report to persistent storage under /mnt/user-data/outputs/ without explicitly warning the user that a file will be created and retained. Persisting an audit of conversation history can compound privacy exposure because it creates a new artifact that may be discoverable or shared later.

VirusTotal

64/64 vendors flagged this skill as clean.