ClawHub
Back to skill

Security audit

Product Tour Capture

Security checks for vulnerabilities and agentic risk

Overview

This skill is a product-tour builder that captures browser screenshots and deploys a tour page, with no malicious behavior found but some expectations and privacy caveats to review.

Install only if you want an agent to modify a Next.js repository, create persistent tour assets, and potentially deploy the result. Use approved demo or test URLs, review screenshots for sensitive content before publishing, and require explicit approval before any Vercel deployment.

A.I.G

Vulnerability Patterns
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs

SkillSpector

Vulnerability Patterns
  • 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
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (5)

Intent-Code Divergence

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The manifest advertises video segment capture and screen recording, while the body actually describes screenshot capture, code generation, route creation, and deployment into the target app. This mismatch can mislead operators into granting the skill broader filesystem, repository modification, deployment, and network access than they intended, creating change-management and trust-boundary risks.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

The skill performs browser automation against live web apps with outbound network access, but it lacks a clear privacy and environment-safety warning. This can lead operators to run it against authenticated, staging, or sensitive environments where screenshots may capture secrets, personal data, or unreleased product information that then get stored in the repo or deployed publicly.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
98% confidence
Finding

The skill claims to capture product-tour video segments, but the operative behavior is different: it captures screenshots, generates a slideshow page, and deploys code changes. That discrepancy can cause users to invoke the skill under false assumptions, increasing the risk of unauthorized repo modification, unexpected publishing, and accidental exposure of application state or internal environments.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The skill modifies a target Next.js repository and deploys the resulting application, but the description does not clearly warn users that it will write files and publish changes. In agentic workflows this is dangerous because users may authorize a seemingly passive capture task that actually performs code generation and deployment, leading to unintended production changes or exposure of internal content.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
70% confidence
Finding

npx commands without a version suffix (e.g. @1.0.0) create a rug-pull risk if the upstream server is compromised and publishes a malicious update.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.