Security audit

guaikei·小红书竞品内容获取

Security checks across malware telemetry and agentic risk

Overview

This skill has a coherent Xiaohongshu research purpose, but it automatically saves full results locally and sends user inputs plus an API token to a third-party service with limited user control.

Review this before installing if your Xiaohongshu searches, competitor targets, note links, or API token are sensitive. Use it only when you are comfortable sending those inputs to guaikei.com and having full result files saved under the skill's local logs directory; delete logs or restrict filesystem access if needed.

SkillSpector

By NVIDIA

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
MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands

Findings (8)

Context-Inappropriate Capability

Low

Confidence: 85% confidence
Finding: The skill writes the full fetched results to a local JSON log file without user opt-in or minimization. Even if the source content is public, local persistence can unintentionally retain scraped content, metadata, and queried URLs on disk, creating privacy and data-handling risk beyond transient retrieval.

Description-Behavior Mismatch

Medium

Confidence: 92% confidence
Finding: The CLI writes the full search output to a local file, and the filename itself embeds the user's keyword. That creates unintended data persistence for a skill described as search/query oriented, and can leak sensitive interests or search terms to other local users, backup systems, or logs without clear disclosure or retention controls.

Vague Triggers

Medium

Confidence: 93% confidence
Finding: The activation text is broad enough that the skill may be invoked whenever a user merely mentions Xiaohongshu and wants to 'understand content,' even without clearly intending third-party data retrieval. Because the skill sends keywords and links to an external API, overbroad triggering can cause unnecessary disclosure of user-supplied data and unintended tool execution.

Missing User Warnings

Medium

Confidence: 96% confidence
Finding: The document mentions dependency on a third-party API only near the end under compliance notes, but the operational instructions do not clearly warn that user-provided Xiaohongshu links, keywords, and related query inputs will be transmitted to guaikei.com during normal use. Users or calling agents may therefore disclose URLs, business research targets, or other sensitive context to an external processor without meaningful notice.

Missing User Warnings

Medium

Confidence: 91% confidence
Finding: The README states that all task results are automatically saved to a local `logs/` directory, but it does not clearly warn users about data persistence, retention, or the sensitivity of the saved content. Because this skill collects scraped competitor notes, comments, and profile data, automatic storage can expose potentially sensitive business intelligence or personal data to other local users, backups, or downstream tooling without the operator realizing it.

Missing User Warnings

Medium

Confidence: 91% confidence
Finding: The code sends both the API token and a user-supplied URL in a GET request, placing sensitive and user-provided data into request parameters. GET parameters are commonly exposed in logs, browser history, proxies, monitoring systems, and upstream infrastructure, increasing the chance of token leakage and unnecessary disclosure of the user's requested target URL.

Missing User Warnings

Low

Confidence: 84% confidence
Finding: The file writes results to local storage without any explicit disclosure in the CLI flow, which can surprise users and leave data artifacts on shared systems. While the impact is limited, undisclosed persistence increases privacy and operational risk, especially if results include account metadata or large scraped datasets.

Missing User Warnings

Medium

Confidence: 94% confidence
Finding: The code logs search activity and results to disk without any warning in this file, and the stored filename contains the user's raw query. Search queries can reveal sensitive interests, plans, health topics, or other private information, so silent local persistence increases privacy and exposure risk.

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:17