ClawHub
Back to skill

Security audit

Code Delegate

Security checks across malware telemetry and agentic risk

Overview

This skill is a code-delegation helper, but it makes strong safety claims about write protection and parallel task control that the artifact does not actually enforce.

Review this before installing if you expect the skill to enforce write protection or safely manage parallel coding agents. Treat the write-guard examples as configuration text, not proven protection, and only use API keys through your normal secret-management process with narrowly scoped credentials.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (3)

Intent-Code Divergence

Medium
Confidence
95% confidence
Finding
The skill repeatedly presents write protection and permission governance as supported capabilities, but the file only shows sample JSON and shell commands that create configuration files. There is no actual enforcement mechanism in the skill content, so users may rely on nonexistent safeguards while delegating code changes with exec access, which can lead to unintended writes to sensitive paths.

Intent-Code Divergence

Low
Confidence
89% confidence
Finding
The FAQ assures users that write protection will not interfere with normal development because project directories are allowed, but no real protection logic is defined anywhere in the skill. This creates a false sense of safety about both what is protected and what is writable, increasing operational risk.

Missing User Warnings

Medium
Confidence
87% confidence
Finding
The skill instructs users to export API keys directly into environment variables without any warning about secret handling, shell history exposure, CI masking, or avoiding accidental logging. In an agent skill focused on code delegation and exec usage, this increases the chance that credentials are mishandled, echoed, committed, or exposed to subprocesses.

VirusTotal

62/62 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.