Security audit

企业安全开发套件

Security checks for vulnerabilities and agentic risk

Overview

The skill is disclosed as an enterprise security bundle, but it includes concrete anti-detection browser automation guidance for avoiding bans and bot detection, so it needs Review before installation.

Install only if you have a clearly authorized security or QA workflow and are comfortable granting the skill read/write/exec access plus access to local secrets, browser profiles, cookies, and API credentials. Avoid using the anti-detection browser guidance against third-party platforms or accounts where automation is not explicitly permitted.

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 (5)

Vague Triggers

Medium

Confidence: 92% confidence
Finding: The trigger table uses broad terms like '认证', '授权', '安全开发', 'Bug', and '调试', which are common across many normal development conversations. In an agent setting, this can cause over-invocation of a high-capability skill that has read/write/exec access, increasing the chance of unintended command execution or sensitive workflow activation without clear user intent.

Vague Triggers

Medium

Confidence: 89% confidence
Finding: The invocation guidance says to use the skill whenever the user needs related functionality, while also advertising broad enterprise security coverage. This ambiguity makes it easier for the agent to route ordinary security or development requests into a powerful composite skill with execution and file-writing capability, creating avoidable safety and privacy risks.

Missing User Warnings

Medium

Confidence: 95% confidence
Finding: The skill advertises file writing, API integration, and command execution as core features but does not pair them with a clear user-facing warning about system changes, outbound data flow, or side effects. In a composite skill with exec/write tools, that omission can lead to silent modification of the environment or transmission of sensitive data beyond what the user expected.

Ssd 4

Medium

Confidence: 97% confidence
Finding: This workflow operationalizes multi-account browser automation with fingerprint isolation, cookie separation, crash recovery, and direct CDP use specifically to improve anti-detection effectiveness. Even with superficial compliance language elsewhere, the concrete guidance materially enables evasion of platform bot detection and can be used for abusive account farming, stealth scraping, or fraud at scale.

Ssd 2

Medium

Confidence: 96% confidence
Finding: The troubleshooting section advises executing anti-detection JavaScript, using direct WebSocket CDP channels, warming cookies, and tuning behavior to avoid platform bans. These are concrete evasion techniques framed as operational advice, which lowers the barrier to bypassing bot-detection controls and facilitates unauthorized automation against third-party platforms.

Static analysis

No suspicious patterns detected.