Security audit

Cdp Browser Pilot

Security checks for vulnerabilities and agentic risk

Overview

This browser-automation skill is disclosed, but it teaches anti-bot evasion and reading protected browser session cookies, so it needs human review before installation.

Install only if you have a legitimate, authorized need for browser automation and are comfortable granting local browser-session access. Avoid using it to bypass site protections or read/export session cookies; those cookies can act like account credentials.

SkillSpector

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
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

Findings (7)

Context-Inappropriate Capability

Medium

Confidence: 96% confidence
Finding: The skill explicitly advertises anti-detection and anti-automation evasion features as a core capability, which goes beyond ordinary browser automation and materially increases abuse potential. In context, these features are paired with platform-specific scraping guidance, making the skill more suitable for bypassing site defenses than for legitimate testing alone.

Context-Inappropriate Capability

High

Confidence: 99% confidence
Finding: The skill documents reading HttpOnly cookies directly from browser profile/SQLite files, which enables access to authenticated session material specifically protected from normal script access. In a generic automation skill, this is a credential-access capability that can facilitate session hijacking, account takeover, and unauthorized data access.

Vague Triggers

Medium

Confidence: 77% confidence
Finding: The description uses broad activation language tied to common productivity and efficiency tasks, which can cause the skill to be invoked in many contexts beyond narrowly justified browser automation. Because the skill also includes sensitive capabilities like anti-detection and cookie extraction, overly broad routing increases the chance of unsafe or unintended use.

Missing User Warnings

High

Confidence: 98% confidence
Finding: The markdown normalizes reading HttpOnly cookies from browser files without a strong warning that this accesses sensitive authenticated data. That omission lowers operator caution around highly privileged session artifacts and makes misuse more likely.

Ssd 2

High

Confidence: 99% confidence
Finding: This section gives concrete anti-detection tactics such as masking webdriver, spoofing plugins/languages, randomizing behavior, and human-like interaction patterns. These are explicit evasion techniques intended to circumvent website automation controls, which materially increase the likelihood of abusive scraping or unauthorized automated access.

Ssd 2

High

Confidence: 95% confidence
Finding: The platform-specific guidance for Xiaohongshu directly discusses strong anti-automation defenses and recommends enabling anti-detection before interacting with the site. That context turns otherwise generic automation into targeted instructions for working around a named platform's defenses.

Ssd 3

High

Confidence: 99% confidence
Finding: The instructions explicitly describe extracting HttpOnly cookies from browser files, which can expose protected session tokens and other sensitive authentication material. Because HttpOnly exists to prevent ordinary script access, documenting file-based extraction is especially dangerous and closely aligned with credential theft tradecraft.

Static analysis

No suspicious patterns detected.