ClawHub
Back to skill

Security audit

HTML Code Painting

Security checks for vulnerabilities and agentic risk

Overview

This skill is a coherent local HTML artwork helper with expected code-generation and preview utilities, and I found no hidden persistence, credential access, network exfiltration, or destructive behavior.

Install this if you want agents to create local, self-contained HTML artwork. Review generated HTML before opening it when the prompt or source material is untrusted, and avoid placing arbitrary third-party HTML into the comparison template.

SkillSpector

Vulnerability Patterns
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (3)

Context-Inappropriate Capability

Medium
Confidence
96% confidence
Finding
The template inserts attacker-controlled markup into multiple containers via innerHTML, so any supplied artwork HTML can include script execution, event handlers, iframes, network requests, or UI overlays. In this skill’s context, the placeholder is explicitly meant to be replaced with arbitrary user- or model-generated HTML, which makes this an actual code-injection sink rather than a theoretical issue.

Vague Triggers

Medium
Confidence
89% confidence
Finding
The activation description is broad enough to capture many generic art or image-related requests, which can cause the agent to invoke this skill outside its intended scope. Over-broad routing is a security-relevant quality issue because it can override safer or more appropriate tools, increase prompt-surface exposure, and make downstream behavior less predictable.

Natural-Language Policy Violations

Medium
Confidence
82% confidence
Finding
The skill metadata is written to enforce Chinese-language behavior without saying that language should follow user preference. This can cause unintended language switching, weaken instruction hierarchy alignment, and reduce usability or policy compliance when the user expects another locale.

Static analysis

No suspicious patterns detected.