Security audit

guaikei-douyin-trending-to-rank-report

Security checks across malware telemetry and agentic risk

Overview

The skill’s Douyin data-collection purpose is mostly coherent, but broad auto-trigger rules and default local storage of collected search, account, and comment data warrant Review before installation.

Install only if you want an API-backed Douyin research tool that can send your search terms, Douyin URLs, limits, and token-authenticated requests to guaikei.com, and that will save full JSON results locally under logs/. Use it for explicit Douyin tasks, review or delete saved logs when they may contain sensitive research topics or public user/comment data, and avoid relying on the broad trigger wording for ambiguous trend or competitor-research requests.

SkillSpector

By NVIDIA

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
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
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access

Findings (8)

Description-Behavior Mismatch

Medium

Confidence: 92% confidence
Finding: The schema explicitly documents `play_addr` as both a playback URL and a video download address, which conflicts with the skill's stated scope excluding download-related use. That mismatch can enable downstream agents or users to repurpose the skill for content downloading, weakening policy controls and increasing the risk of misuse around unauthorized media retrieval.

Vague Triggers

Medium

Confidence: 91% confidence
Finding: The skill defines a broad '泛化触发' rule that activates on generic short-video research or multi-step investigation requests even when the user does not explicitly mention Douyin. This can cause the agent to invoke external data collection unexpectedly, broadening the skill's reach beyond clear user intent and increasing the risk of over-collection, platform mismatch, or unintended third-party API use in unrelated workflows.

Vague Triggers

Medium

Confidence: 91% confidence
Finding: The README maps very broad natural-language phrases such as '搜索 / 搜一下 / 找' to command execution, which can cause unintended activation from ordinary user phrasing. In an agent setting, this raises the risk of the skill running data-collection actions when the user did not explicitly intend to invoke Douyin scraping, leading to unnecessary external requests and possible collection or persistence of third-party data.

Vague Triggers

Medium

Confidence: 93% confidence
Finding: The intent-recognition rule prioritizes broad hot-list triggers like '热搜 / 热点 / 榜单 / 今天什么火', which overlap heavily with common conversation and could invoke the skill unexpectedly. Because this skill performs external data retrieval and logging, accidental activation can cause unintended network access, data processing, and storage beyond user expectations.

Missing User Warnings

Medium

Confidence: 89% confidence
Finding: The README advertises automatic JSON log export to a local logs directory but does not clearly warn that outputs may contain sensitive or regulated content, such as scraped comments, account identifiers, or query history. Persistent logging increases exposure through local retention, accidental sharing, backups, or reuse by other tools, especially in a skill designed for bulk collection and analysis.

Missing User Warnings

Low

Confidence: 74% confidence
Finding: The script silently persists fetched comment data to a local JSON file while only printing generic success messages. Because comment datasets can contain personal data or sensitive research material, undisclosed local storage can create privacy, retention, and accidental disclosure risk on shared systems or in automated environments.

Missing User Warnings

Medium

Confidence: 93% confidence
Finding: The CLI persists the full fetched results to a local JSON file after completing the request, but this file gives no explicit warning or opt-in for local retention. Because the skill handles creator/post data and may be run in shared or automated environments, silent disk persistence can expose collected data to other local users, backup systems, or later unintended reuse.

Missing User Warnings

Medium

Confidence: 93% confidence
Finding: The CLI persists full search output to a local JSON file, including the queried keyword and returned results, without obtaining user consent or clearly disclosing that data will be stored on disk. In this skill’s context, search terms and collected Douyin results may contain sensitive research topics, user/account data, or comment-derived content, so silent persistence increases the risk of unintended local data exposure to other users, backup systems, or log collection tooling.

VirusTotal

VirusTotal findings are pending for this skill version.

View on VirusTotal

Static analysis

No suspicious patterns detected.