虾安全

• Outcome: Pass
• Latest audit: May 29, 2026 at 8:42 PM UTC
• Version:
  1.0.0

Overview

The skill appears to be a local audit tool whose persistence is purpose-aligned, but it may leave audit history on disk that users should treat as sensitive.

Install only if you are comfortable with the skill keeping local audit snapshots and alert logs. On shared machines or systems with broad backups, review or periodically delete those files and restrict access to the audit directory.

SkillSpector

By NVIDIA

Vulnerability Patterns

• 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
• Excessive Agency: Unrestricted Tool Access, Autonomous Decision Making, Scope Creep

Findings (2)

Missing User Warnings

Medium

• Confidence: 91% confidence
• Finding: The script saves full audit results to timestamped snapshot files under the user's home directory. Those results can contain sensitive metadata about the agent workspace, including filenames, findings, and potentially excerpts-derived indicators about identity, memory, and installed skills, creating a privacy and persistence risk if other local processes or users can read them.

Missing User Warnings

Medium

• Confidence: 88% confidence
• Finding: The alert logger appends security findings to a persistent log file without any explicit consent or notice. Repeated logging can accumulate sensitive operational details about the user's workspace and detected issues, which may later be exposed through local compromise, backups, or shared environments.

VirusTotal

66/66 vendors flagged this skill as clean.