Security audit

guaikei-douyin-data-grabber

Security checks across malware telemetry and agentic risk

Overview

The skill is a coherent Douyin data tool, but it needs Review because it can trigger too broadly and automatically stores large scraped datasets containing user/comment identifiers.

Install only if you are comfortable sending Douyin queries and your GUAIKEI_API_TOKEN to guaikei.com, and only use it for explicit Douyin tasks. Protect or delete the generated logs, because they can contain usernames, IDs, comments, profile links, and IP-region labels, and verify your legal, platform, and internal data-handling obligations before bulk comment collection.

SkillSpector

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 (10)

Description-Behavior Mismatch

Medium

Confidence: 92% confidence
Finding: The CLI persists the full fetched Douyin results to a local JSON file after printing them, which creates an additional data-retention channel beyond the obvious stdout behavior. Because the skill handles scraped social-media content and potentially comments or account data, silent local storage increases privacy, leakage, and unintended reuse risk, especially on shared systems or automated runners.

Vague Triggers

High

Confidence: 94% confidence
Finding: The skill instructs invocation for broad, implicit scenarios such as general short-video research or multi-step workflows even when the user never mentioned Douyin. That can cause the agent to route user requests to an external data-collection skill without clear user consent or platform specificity, leading to unintended third-party queries and privacy/expectation violations.

Natural-Language Policy Violations

Medium

Confidence: 83% confidence
Finding: Forcing a Chinese-only interface without user opt-in can mislead or exclude users and may cause the agent to present outputs in an unexpected language. In a multi-user or multilingual environment, this undermines informed use and can contribute to incorrect interpretation of results or commands.

Vague Triggers

Medium

Confidence: 91% confidence
Finding: The README teaches the agent to trigger this skill from broad, everyday phrases like '搜索', '找', or topical requests embedded in larger workflows, which can cause over-invocation on ambiguous user intent. In a data-grabbing skill, that increases the chance of unintended scraping, unnecessary third-party API calls, and collection of public-but-still-sensitive comment or account data without clear user consent.

Vague Triggers

Medium

Confidence: 94% confidence
Finding: The intent-recognition priority rules rely on generic terms such as '热搜', '搜索', '评论', and '视频' with limited scope constraints, making false matches likely in normal conversation. Because this skill performs external data retrieval and logging, ambiguous routing can lead to unintended data access, API-token use, and persistence of scraped results when the user may have meant a different platform or a non-scraping task.

Missing User Warnings

Medium

Confidence: 88% confidence
Finding: The README emphasizes automatic JSON log generation and later states that search, post, and comment results are saved under a logs directory, but it does not warn that these files may contain scraped usernames, comments, account identifiers, or other personal data. This creates a real risk of silent local retention, over-collection, and secondary exposure if logs are copied, shared, committed to source control, or left on multi-user systems.

Missing User Warnings

Medium

Confidence: 89% confidence
Finding: The document explicitly enables bulk retrieval of Douyin comment data but provides no privacy, consent, retention, or platform-terms warning. In the context of a skill marketed for competitor research and public-opinion analysis, this omission can encourage collection and downstream analysis of user-generated data at scale without appropriate safeguards, increasing privacy and compliance risk.

Missing User Warnings

Medium

Confidence: 94% confidence
Finding: The code writes result data to a local JSON file without any user-facing notice or consent in this file, creating hidden persistence of scraped data. This is risky because operators may assume the tool only fetches and prints results, while the local file can later be accessed by other users, backup systems, or unrelated processes.

Missing User Warnings

Medium

Confidence: 79% confidence
Finding: The CLI writes full search results to a local JSON file automatically, but the user is not explicitly warned or asked to opt in. Because search results may contain operational research data, account information, or comment content, this can leave sensitive data on disk where other local users, backup systems, or later processes can access it unexpectedly.

Natural-Language Policy Violations

Medium

Confidence: 93% confidence
Finding: The authentication failure message embeds a direct personal contact channel ('微信: 13395823479') in application error handling. This can socially steer users toward off-platform support, facilitate phishing or impersonation if logs/screenshots are shared, and unnecessarily exposes personal contact data in runtime messages for an auth path that may be triggered broadly.

VirusTotal

VirusTotal findings are pending for this skill version.

View on VirusTotal

Static analysis

No suspicious patterns detected.