ClawHub
Back to skill

Security audit

Skill Starter Pack

Security checks across malware telemetry and agentic risk

Overview

This skill does what it says, but it can bulk-install many unpinned skills that materially change an agent’s behavior.

Review the exact skills before installing, especially Developer or Full. Prefer starting with Essential, verify publishers and versions where possible, and understand how to disable or uninstall memory, browser automation, GitHub, and proactive-agent skills before enabling them.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • 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 (2)

Lp3

Medium
Category
MCP Least Privilege
Confidence
95% confidence
Finding
The skill instructs the agent to execute a shell-based installer (`python <skill-dir>/scripts/install_pack.py ...`) and to invoke external tooling (`clawhub` or `npx clawhub`), but the skill does not declare shell permissions. This is a real security issue because users and enforcement layers cannot accurately understand or constrain the skill's execution capabilities, and the script can install multiple third-party skills from the network, expanding trust to additional unreviewed code.

Description-Behavior Mismatch

Medium
Confidence
94% confidence
Finding
The script is presented as a curated starter-pack installer, but it silently expands scope by accepting arbitrary extra skill names from the command line and installing them. In a security-sensitive agent ecosystem, that bypasses the safety expectation of 'curated only' and can lead users to install unreviewed or malicious skills under the trusted banner of the starter pack.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.