Skill Install Guard｜技能安装守门员

• Outcome: Pass
• Latest audit: May 29, 2026 at 8:36 PM UTC
• Version:
  0.4.3

Overview

This skill is a disclosed install-review helper that can run a user-provided install command, so it is useful but should be used carefully.

Use dry-run or stop-before-install first, review the generated report, and only provide a simple trusted install command such as a known ClawHub install invocation. Avoid privileged commands or shell interpreters, set an expected directory, and treat the tool as review support rather than a guarantee that a target skill is safe.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Behavioral AST: exec() Call, eval() Call, Dynamic Import
• 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)

subprocess module call

Medium

• Category: Dangerous Code Execution

Content

"Install command must be a direct executable invocation without shell operators: "
            + ", ".join(sorted(set(bad)))
        )
    return subprocess.run(argv, cwd=str(cwd) if cwd else None, text=True, capture_output=True)


def maybe_json(text: str):

• Confidence: 88% confidence
• Finding: return subprocess.run(argv, cwd=str(cwd) if cwd else None, text=True, capture_output=True)

VirusTotal

64/64 vendors flagged this skill as clean.