ClawHub
Back to skill

Security audit

Ai Agent Helper Free

Security checks across malware telemetry and agentic risk

Overview

This skill is mostly a prompt-design guide, but it requests shell execution authority that is broader than its stated purpose and includes network troubleshooting commands.

Review this skill before installing if your agent environment grants declared tools automatically. It appears non-destructive and educational, but users should consider removing or denying exec access unless they specifically want the agent to run local/network troubleshooting commands.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • 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 (2)

Context-Inappropriate Capability

Medium
Confidence
92% confidence
Finding
The skill declares the `exec` tool even though its stated purpose is prompt engineering and ReAct design assistance, not system administration or network diagnostics. Exposing generic command execution broadens the attack surface and can enable unsafe command execution if the agent is induced to follow troubleshooting or user-supplied instructions through the shell.

Context-Inappropriate Capability

Medium
Confidence
95% confidence
Finding
The troubleshooting guidance explicitly instructs execution of `ping` and inspection of firewall/proxy settings, which introduces operational system/network behavior outside the skill's design-helper scope. In an agent environment with `exec` available, such instructions can be turned into real command execution, creating a pathway from benign documentation to unintended shell actions and environment probing.

VirusTotal

VirusTotal findings are pending for this skill version.

View on VirusTotal

Static analysis

No suspicious patterns detected.