Skill Proposal Gen

• Outcome: Review
• Latest audit: May 29, 2026 at 9:33 PM UTC
• Version:
  1.0.1

Overview

This appears to be a local proposal-generation skill, but it can read or write JSON and Markdown files outside its intended folder if given a crafted project name.

Review before installing. Avoid project names containing slashes, backslashes, absolute paths, or '..'. The skill stores proposal content locally and should be updated to validate project names and force all reads and writes to remain inside its proposals folder.

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: 83% confidence
• Finding: The skill is documented with extremely broad purpose statements such as '基础功能实现', '状态管理', and '数据处理', plus a generic execute(action="connect") example without clear trigger boundaries or user-consent conditions. In an agent ecosystem, vague scope increases the chance of over-broad invocation, misuse in unintended workflows, and unsafe assumptions about what operations the skill may perform.

VirusTotal

67/67 vendors flagged this skill as clean.