ClawHub
Back to skill

Security audit

Code Dev V1 Tool Free

Security checks across malware telemetry and agentic risk

Overview

This skill is a local structured coding workflow helper whose file writes and command use are mostly disclosed and aligned with its purpose.

Before installing, be aware that this skill can guide an agent to read project files, run local shell commands, run tests, and create retained files under ~/code for preferences and checkpoints. Use it for personal coding workflows, review any command before execution, and periodically inspect or clean ~/code/memory.md and ~/code/checkpoints if the project contains sensitive details.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (3)

Intent-Code Divergence

Medium
Confidence
94% confidence
Finding
The skill claims it does not automatically execute code, yet elsewhere explicitly allows and demonstrates use of exec for shell commands, initialization, and running tests. This mismatch can mislead users and agents about the real execution boundary, causing unreviewed command execution and unsafe trust in the skill's safety guarantees.

Vague Triggers

Medium
Confidence
87% confidence
Finding
The skill advertises an unusually broad set of trigger keywords and capability coverage, which increases the chance the agent invokes it in contexts beyond the author's intended scope. Over-broad matching can activate file-writing and command-related behaviors unexpectedly, expanding the attack surface for prompt injection or accidental misuse.

Missing User Warnings

Medium
Confidence
96% confidence
Finding
The skill instructs persistent writes to ~/code/memory.md and checkpoint directories without providing an upfront, prominent disclosure that local files will be created and retained. Silent persistence creates privacy and integrity risks, especially if users do not realize their preferences, task history, or potentially sensitive project details are being stored on disk.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.