Security audit

浏览器自动化(免费版)

Security checks for vulnerabilities and agentic risk

Overview

This browser automation skill is mostly aligned with its purpose, but it needs review because it encourages credentialed form submission and external callbacks without clear user-control safeguards.

Review this skill before installing. Use it only for sites and forms you intentionally authorize, avoid giving it real passwords unless the agent asks for explicit confirmation and handles secrets safely, and treat callback URLs as external data sharing destinations that must be trusted.

SkillSpector

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (4)

Vague Triggers

Medium

Confidence: 77% confidence
Finding: The activation/scope text is excessively broad and maps the skill to generic efficiency, automation, batch processing, and workflow optimization requests beyond its narrowly claimed feature set. In an agent ecosystem, overbroad routing increases the chance the skill is invoked for sensitive browsing, scraping, or form-submission tasks without adequate user intent checks or safety warnings.

Missing User Warnings

High

Confidence: 96% confidence
Finding: The skill includes a login form autofill and submission example using credentials, but provides no warning about secrets handling, account actions, or side effects. In an agent context, this can lead to automated credential use, unintended logins, or submission of sensitive data to third-party sites without an explicit confirmation boundary.

Missing User Warnings

Medium

Confidence: 88% confidence
Finding: The skill advertises data extraction plus callback URL support without explaining that scraped content or user-supplied data may be sent to external destinations. That omission is dangerous because users may trigger exfiltration of page contents or sensitive inputs to attacker-controlled endpoints under the guise of normal automation.

Missing User Warnings

Medium

Confidence: 74% confidence
Finding: The quick-start commands write output files locally without any warning that the skill modifies the filesystem. In agent-driven environments, undocumented file writes can overwrite existing data, create artifacts in sensitive directories, or violate least-surprise expectations for a read/extract workflow.

Static analysis

No suspicious patterns detected.