baoyu-danger-gemini-web

Warn

Audited by Socket on Feb 15, 2026

1 alert found:

Security
SecurityMEDIUM
SKILL.md

[Skill Scanner] Backtick command substitution detected All findings: [HIGH] command_injection: Backtick command substitution detected (CI003) [AITech 9.1.4] [HIGH] command_injection: Backtick command substitution detected (CI003) [AITech 9.1.4] The package is a local CLI wrapper that uses a reverse‑engineered Gemini Web interface and reuses browser cookies/profile to authenticate. The manifest itself contains no explicit malicious code or obfuscation, but the design carries elevated risk: reading browser cookies and profile data and storing prompts/sessions locally can expose sensitive credentials or user content if misused or misconfigured. Before use, review the implementation files for exact network endpoints and any non‑Google destinations, restrict authentication to an isolated Chrome profile, and protect or rotate stored cookies and session files. Treat the tool as higher‑risk than official API clients until its implementation is audited. LLM verification: SUSPICIOUS: The skill’s described consent flow and reliance on a reverse-engineered Gemini Web API, plus local credential/cookie handling, deviate from standard, official API usage. While there is no explicit malicious code shown, the design elevates privacy and policy concerns (consent management, cookies, local data). If used as intended, it could function with user consent; however, the footprint (consent file, cookies, browser auth) is broader than a typical, officially-supported integration

Confidence: 75%Severity: 75%
Audit Metadata
Analyzed At
Feb 15, 2026, 05:14 AM
Package URL
pkg:socket/skills-sh/xuanxuan1983%2Fbaoyu-xuanyi-skills%2Fbaoyu-danger-gemini-web%2F@3dc55c0dd2bc7a5d63ab6224f7619aab0bd4beec