Security audit

浏览器代理(专业版)

Security checks across malware telemetry and agentic risk

Overview

This browser automation skill is transparent about its features, but it teaches stealth scraping, proxy rotation, fingerprint spoofing, and session persistence in ways users should review carefully before installing.

Install only if you will use it on sites and accounts you are authorized to automate. Avoid using the anti-detection, residential proxy, or fingerprint-spoofing guidance to bypass access controls, rate limits, paywalls, or site rules. Treat saved cookies and sessions as credentials: keep them encrypted, out of version control, permission-restricted, scoped per site, and deleted when no longer needed.

SkillSpector

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access

Findings (7)

Intent-Code Divergence

Medium

Confidence: 93% confidence
Finding: The skill claims it does not solve CAPTCHAs, but elsewhere explicitly promotes anti-detection, fingerprint spoofing, proxy rotation, and human-behavior simulation to avoid anti-bot systems. That combination still enables circumvention of site defenses and creates a misleading safety framing that may normalize abusive scraping against protected services.

Missing User Warnings

Medium

Confidence: 89% confidence
Finding: The skill documents persistent storage of authenticated cookies, localStorage, and sessionStorage in local JSON files, which are highly sensitive because they can allow session hijacking if accessed by another process or user. Although later FAQ text mentions permissions and encryption, the core feature description lacks prominent warnings, secure-by-default handling requirements, and clear consent boundaries for storing active sessions.

Missing User Warnings

Medium

Confidence: 81% confidence
Finding: The skill describes saving CSVs, screenshots, reports, cookies, and performing automatic cleanup, but does not consistently warn users about where files are written, how long they are retained, or what deletion behavior may occur. In automation contexts this can lead to unintended storage of sensitive business or personal data and accidental destructive actions on local files.

Missing User Warnings

Medium

Confidence: 93% confidence
Finding: The documentation explicitly enables session persistence and cookie storage for browser automation but provides no warning that these artifacts may contain active authentication tokens or other sensitive account data. In the context of a browser agent with scraping and stealth capabilities, persisted cookies materially increase the risk of credential/session theft, unintended reuse across tasks, and insecure storage of secrets on disk.

Missing User Warnings

Medium

Confidence: 88% confidence
Finding: The example references proxy credentials through environment variables, which is better than hardcoding, but the documentation still omits any warning that these values are secrets that can leak through logs, shell history, process inspection, or misconfigured CI environments. Because the skill also supports proxy rotation and stealth browsing, compromised proxy credentials could enable unauthorized traffic routing, abuse of paid proxy infrastructure, or attribution masking.

Ssd 4

Medium

Confidence: 96% confidence
Finding: This section repeatedly operationalizes stealth mechanisms such as UA rotation, fingerprint spoofing, proxy usage, randomized delays, and human-like interaction as standard workflow features. In the context of a browser automation skill with exec capability, this meaningfully increases the likelihood of unauthorized scraping and deliberate evasion of website defenses.

Ssd 4

Medium

Confidence: 95% confidence
Finding: The examples and optimization advice form a practical playbook for scaling evasive collection: rotating residential proxies, tuning request cadence, maintaining fingerprint consistency, and collecting dynamic or paginated content at volume. This materially lowers the barrier to abuse by translating abstract evasion into actionable operational steps.

VirusTotal

VirusTotal findings are pending for this skill version.

View on VirusTotal

Static analysis

No suspicious patterns detected.