guaikei·小红书博主作品

• Outcome: Review
• Latest audit: Aug 13, 2026 · 4:15 AM UTC
• Version: 1.0.0

Overview

The skill is a public Xiaohongshu data lookup tool that uses guaikei.com and saves results locally, with the main risks disclosed enough for a benign verdict.

Install only if you are comfortable sending Xiaohongshu keywords, links, and fetched public data to guaikei.com and keeping JSON result logs on the local machine. Keep GUAIKEI_API_TOKEN private, clear artifact/logs when results are no longer needed, and avoid using the tool for private, login-only, or legally restricted content.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Trigger Abuse: Overly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
• MCP Tool Poisoning: Hidden Instructions, Unicode Deception, Parameter Description Injection
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access

Findings (7)

Description-Behavior Mismatch

Medium

• Confidence: 95% confidence
• Finding: The README presents a much broader data-mining and monitoring capability set than the manifest’s stated narrow purpose of routing Xiaohongshu links for content understanding. This scope mismatch is dangerous because users and platform operators may grant trust or permissions based on the manifest while the documentation encourages broader scraping, bulk collection, and competitive intelligence workflows that materially change the tool’s risk profile.

Intent-Code Divergence

Low

• Confidence: 84% confidence
• Finding: The documentation claims the tool does not involve user privacy data, yet it also advertises comment retrieval and returning broad public data. Even if the data is publicly accessible, comments, profile-linked activity, and aggregated engagement data can still constitute personal data or sensitive behavioral information in some jurisdictions, so the statement may mislead users about privacy exposure.

Description-Behavior Mismatch

High

• Confidence: 89% confidence
• Finding: This CLI adds keyword-based Xiaohongshu search functionality that is broader than the skill metadata, which states the skill should only route and analyze provided Xiaohongshu links. That scope expansion can cause the agent to collect or surface unrelated platform content a user did not explicitly provide, increasing privacy, policy, and least-privilege risk in an agent environment.

Missing User Warnings

Medium

• Confidence: 95% confidence
• Finding: The skill instructs the agent to send user-provided Xiaohongshu links and retrieved content through a third-party API, but the pre-invocation description does not clearly disclose that external transfer will occur. This can cause users to unknowingly share URLs, query terms, and potentially sensitive research targets with an outside service, creating privacy, confidentiality, and consent risks.

Missing User Warnings

Medium

• Confidence: 92% confidence
• Finding: The README states that all task results are automatically saved to a local logs/ directory, but it does not clearly warn that fetched content, profile URLs, note URLs, search keywords, and analysis results may be written to disk. This can create unintended local data retention, exposing scraped content and user-supplied links to other local users, backups, or later exfiltration if the host is compromised.

Missing User Warnings

Medium

• Confidence: 80% confidence
• Finding: The tool persists fetched results to a local JSON file automatically without explicit user notice or opt-in. Because Xiaohongshu URLs may contain query parameters like xsec_token and the fetched results may include profile/content metadata, this can leave sensitive or privacy-relevant data on disk where other local users, processes, backups, or logs may access it unexpectedly.

Missing User Warnings

Medium

• Confidence: 86% confidence
• Finding: The code persists full search output to a local JSON file named with the user keyword, without user notice or consent in this file. Search terms and returned content may contain sensitive or personal data, and writing them to disk increases exposure through local compromise, log collection, backups, or multi-user system access.

VirusTotal

61/61 vendors flagged this skill as clean.

Static analysis

Detected: suspicious.exposed_secret_literal

File appears to expose a hardcoded API secret or token.

Critical

• Code: suspicious.exposed_secret_literal
• Location: SKILL.md:16