ClawHub
Back to skill

Security audit

Claude Code Sdk

Security checks across malware telemetry and agentic risk

Overview

This skill is a coherent Claude Code wrapper, but it broadly instructs agents to delegate nearly all coding work to an external agent with file editing and shell authority.

Install only if you intentionally want coding tasks delegated to Claude Code with permission to read, edit, write files, run shell commands, and use project/user Claude settings. Use it in a controlled project directory under version control, avoid sensitive repositories unless external Claude Code processing is allowed, start fresh sessions for unrelated work, and treat --log-file as sensitive because it may store code, prompts, or other confidential output.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (1)

Vague Triggers

Medium
Confidence
91% confidence
Finding
The skill description says it should be preferred for any code-programming-related problem, which is an overly broad routing rule for a powerful external coding agent. This can cause unnecessary delegation of diverse development tasks to an autonomous tool, increasing the chance of unsafe code changes, over-collection of context, or bypassing safer, task-specific skills and human review.

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

Detected: suspicious.dangerous_exec

Shell command execution detected (child_process).

Critical
Code
suspicious.dangerous_exec
Location
scripts/run_claude.mjs:10