Security audit

guaikei-xhs-content-collector

Security checks across malware telemetry and agentic risk

Overview

This skill appears to do what it says: fetch public Xiaohongshu data through Guaikei, while sending user queries/URLs to that service and saving successful results locally.

Install only if you are comfortable sending Xiaohongshu keywords, note/profile URLs, and the GUAIKEI_API_TOKEN-authenticated requests to guaikei.com. Avoid submitting sensitive URLs or secrets, and periodically review or delete the generated logs if the collected research targets or results should not remain on disk.

SkillSpector

By NVIDIA

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
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)

Description-Behavior Mismatch

Medium

Confidence: 93% confidence
Finding: The CLI writes full search results to a local JSON file after completing the request, creating persistent storage of collected platform data without an explicit user-facing disclosure or retention control. Even if the data is public, local persistence increases privacy, compliance, and data-handling risk because downstream users may assume the tool only performs transient retrieval.

Missing User Warnings

Medium

Confidence: 96% confidence
Finding: The skill sends user-supplied Xiaohongshu URLs and query parameters to a third-party API service (`guaikei.com`), but the warning is buried later in the document rather than presented prominently at the point of use. This can cause unintended disclosure of user-provided targets, tokens embedded in URLs, or research intent to an external service without sufficiently informed consent.

Missing User Warnings

Medium

Confidence: 83% confidence
Finding: Automatically saving all task results to a local logs/ directory creates unnecessary data persistence and may retain sensitive or regulated content longer than users expect. Even if the source data is public, saved search results, URLs, comments, and derived outputs can expose research intent, monitoring targets, or collected datasets to other local users, backups, or later compromise.

Missing User Warnings

Low

Confidence: 84% confidence
Finding: The code places the API token, URL, and limit in a GET request, which commonly exposes sensitive values in URLs, server logs, browser history, intermediary proxies, and monitoring systems. Even if transport uses HTTPS, query-string secrets are more likely to leak through operational telemetry and referrer-style propagation than values sent in headers or POST bodies.

Missing User Warnings

Medium

Confidence: 91% confidence
Finding: Search results are saved to disk at the write site with no explicit warning or consent prompt, so users may unknowingly leave behind a local corpus of collected content and metadata. This can expose sensitive research topics, operational queries, or scraped public data to other local users, backups, or later compromise of the host.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

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