guaikei-douyin-trend-signal-detector

• Outcome: Review
• Latest audit: Aug 24, 2026 · 3:04 AM UTC
• Version: 1.0.0

Overview

This appears to be a real Douyin public-data tool, but it needs review because it can run on ambiguous research requests, sends token and query data to a third-party API, and saves collected results locally by default.

Install only if you are comfortable sending Douyin search terms, profile or video URLs, and your GUAIKEI_API_TOKEN to guaikei.com, and with search/post/comment results being saved under the skill's logs directory. Use explicit Douyin wording to avoid accidental activation, avoid sensitive monitoring targets on shared machines, and delete logs when they are no longer needed.

SkillSpector

Vulnerability Patterns

• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Excessive Agency: Unrestricted Tool Access, Autonomous Decision Making, Scope Creep
• 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

Findings (11)

Description-Behavior Mismatch

Medium

• Confidence: 88% confidence
• Finding: The changelog describes capabilities that exceed the stated skill scope, including general Douyin search, hot-search querying, bulk retrieval of an author's published works, and comment scraping/analysis. This mismatch is dangerous because downstream agents or users may rely on the broader documented behavior and invoke data-collection functions outside the intended trend-signal use case, increasing the risk of overcollection, privacy issues, and policy/scope bypass.

Description-Behavior Mismatch

Medium

• Confidence: 85% confidence
• Finding: The CLI writes fetched Douyin comment data to a local JSON file after completing the requested retrieval. Because comments may contain personal data or sensitive content, persisting them to disk creates an undeclared data-retention surface and increases exposure through local compromise, shared environments, or later unintended reuse.

Context-Inappropriate Capability

Low

• Confidence: 87% confidence
• Finding: The CLI silently persists fetched results to a local JSON file, creating a secondary copy of potentially sensitive or regulated data without clear necessity or user consent. In an agent/skill context, undisclosed local retention increases privacy, data handling, and multi-tenant leakage risk, especially if logs are accessible to other processes or users.

Description-Behavior Mismatch

Low

• Confidence: 84% confidence
• Finding: The CLI persists the full search output to a local JSON file via log.taskWrite without any clear user-facing disclosure or opt-in. In a data-fetching skill, silent local persistence can expose searched keywords and returned content to other local users, backup systems, or later processes, especially if results contain sensitive research topics or identifiers.

Vague Triggers

Medium

• Confidence: 90% confidence
• Finding: The generalized trigger says the skill should activate even when users do not explicitly mention Douyin, including broad requests like researching a 'track' or competitor analysis embedded in multi-step workflows. That can cause the agent to invoke this skill for unrelated platforms or ambiguous research tasks, leading to unintended data collection, user surprise, and cross-skill misrouting.

Missing User Warnings

Medium

• Confidence: 89% confidence
• Finding: The README describes data collection and export features, but it does not present a prominent up-front warning that routine use will collect Douyin public data and automatically persist structured JSON logs locally. In a skill that handles social-media content and comments, silent local retention can expose scraped content, search targets, and analysis artifacts to unintended users or downstream tools.

Missing User Warnings

Medium

• Confidence: 93% confidence
• Finding: The skill requires a GUAIKEI_API_TOKEN and communicates with an external service, but the README does not clearly warn that user queries and the credential will be transmitted off-box to guaikei-controlled infrastructure. This can mislead operators about trust boundaries and creates credential-handling and data-sharing risk, especially when queries may contain sensitive monitoring targets or business intelligence interests.

Vague Triggers

Medium

• Confidence: 87% confidence
• Finding: The natural-language trigger mapping uses broad everyday phrases such as '搜索 / 搜一下 / 找', which can match common conversational requests without sufficient Douyin-specific context. That increases the chance of unintended invocation, causing external requests, token use, and local logging when the user did not intend to run this data-collection skill.

Vague Triggers

Medium

• Confidence: 84% confidence
• Finding: Keywords like '热搜 / 热点 / 榜单 / 今天什么火' are contextually broad and may activate on generic discussions unrelated to Douyin. In this skill, accidental activation is more concerning because execution can call an external service and generate local logs, expanding the impact beyond a harmless misclassification.

Missing User Warnings

Medium

• Confidence: 90% confidence
• Finding: The file write occurs without any user-facing disclosure in this command, so users may believe the tool only returns results to stdout while it also leaves a durable local artifact. Hidden persistence is dangerous because it broadens data exposure, complicates compliance, and may leak collected content through backups, shared workspaces, or subsequent log processing.

Missing User Warnings

Low

• Confidence: 88% confidence
• Finding: The program writes search results to a local file by default, but the user-facing flow does not clearly warn that execution creates persistent JSON logs. This can violate user expectations and leak search history or result data through filesystem access, especially on shared hosts or managed agent environments where artifacts are retained.

VirusTotal

63/63 vendors flagged this skill as clean.

Static analysis

No suspicious patterns detected.