ClawHub
Back to skill

Security audit

Agent Web Cpu

Security checks for vulnerabilities and agentic risk

Overview

This skill is a disclosed transweb.cn browser automation helper that saves generated Markdown and updates its own app list, with no evidence of hidden, destructive, or unrelated behavior.

Install only if you are comfortable allowing this skill to automate transweb.cn in a browser, run page evaluation to extract results, save generated Markdown files in Downloads, and update its local app registry. Prefer the sandbox browser profile unless you intentionally need your logged-in browser session, and review generated files plus any newly added app entries.

SkillSpector

Vulnerability Patterns
  • 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
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Findings (3)

Missing User Warnings

Medium
Confidence
92% confidence
Finding
The README states that generated content is automatically saved to ~/Downloads without clearly warning the user or indicating consent. In an agent skill that performs browser automation and content generation, silent local file writes can surprise users, create clutter, overwrite expected workflows, or persist sensitive/generated data onto disk where other software or users may access it.

Missing User Warnings

Medium
Confidence
90% confidence
Finding
The skill instructs the agent to automatically create and register applications by modifying apps.json when matching fails, without explicit user confirmation that a local registry file will be changed. This creates an integrity risk because ambiguous or adversarial prompts could silently persist untrusted entries and alter future behavior of the skill.

Missing User Warnings

Medium
Confidence
94% confidence
Finding
The skill saves generated Markdown directly to the user's Downloads folder without an explicit warning or confirmation. Automatic file creation in a user-visible location can be abused to drop misleading, sensitive, or unwanted content and may overwrite expectations about what the agent is allowed to persist locally.

Static analysis

No suspicious patterns detected.