guaikei-douyin-thorough-data-miner

• Outcome: Review
• Latest audit: Aug 25, 2026 · 4:35 AM UTC
• Version: 1.0.0

Overview

This skill mostly matches its Douyin research purpose, but it needs review because it can broadly trigger external scraping and automatically save large result sets locally.

Install only if you are comfortable with a third-party Douyin API provider receiving your token, keywords, and target URLs, and with scraped results being saved under the skill's logs directory by default. Review or narrow activation rules so generic research or trend questions do not automatically use this skill.

SkillSpector

Vulnerability Patterns

• 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
• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access

Findings (11)

Description-Behavior Mismatch

Medium

• Confidence: 94% confidence
• Finding: The schema explicitly documents `play_addr` as both a playback URL and a video download address, which materially expands the skill's apparent capability beyond the stated scope. In a Douyin data-mining skill, exposing a download endpoint can enable unauthorized content acquisition or facilitate watermark-bypass workflows even if the manifest claims such use is unsupported.

Description-Behavior Mismatch

Medium

• Confidence: 89% confidence
• Finding: The CLI persists the full scraped results to a local JSON file after completing the retrieval task, even though the skill is presented primarily as a data-retrieval utility. Stored post metadata can contain sensitive or regulated content, and silent local retention increases the risk of unintended disclosure through shared workspaces, backups, or later exfiltration. In this skill context, the danger is elevated because it processes third-party social-media data at scale, so users may not expect durable storage of scraped output beyond terminal/JSON response delivery.

Description-Behavior Mismatch

Medium

• Confidence: 91% confidence
• Finding: The CLI writes the full search output to a local file whose name includes the user’s keyword, creating undeclared data persistence beyond the expected query behavior. Search terms and returned results may contain sensitive research interests, account data, or other user-derived content, and storing them without explicit consent or retention controls increases privacy and data exposure risk on shared or compromised systems.

Missing User Warnings

Medium

• Confidence: 87% confidence
• Finding: The skill states that outputs are automatically saved to `logs/` with timestamps and query parameters, but it does not prominently warn users that searched keywords, URLs, and collected public-comment data will be written to disk. This can create unintended local data retention, exposing potentially sensitive research topics or regulated content through filesystem access, backups, or shared workspaces.

Vague Triggers

Medium

• Confidence: 91% confidence
• Finding: The natural-language trigger examples are broad enough to match common conversational phrases like '搜索', '找', or '视频', which can cause unintended activation of this skill outside a clearly scoped Douyin data-mining request. In an agent environment, this increases the risk of over-collection, surprising tool execution, and incorrect routing of user intent to a third-party data service.

Vague Triggers

Medium

• Confidence: 89% confidence
• Finding: The competitor-monitoring examples use ambiguous phrases like '查看这个博主的所有作品 / 主页' and '获取他最近50条视频' without requiring clear consent, platform context, or explicit identification that the target is a Douyin public account. This can cause the agent to infer surveillance-style data collection from vague user language and trigger bulk scraping against third-party profiles unintentionally.

Vague Triggers

Medium

• Confidence: 88% confidence
• Finding: The comment-analysis triggers rely on generic terms such as '评论' and '留言', which commonly appear in everyday requests unrelated to Douyin scraping. In this skill, those broad triggers are more dangerous because they map directly to extraction of comment data from a third-party platform, creating a meaningful risk of unintended data retrieval and privacy-invasive analysis.

Vague Triggers

Medium

• Confidence: 90% confidence
• Finding: The hotlist triggers include very broad phrases like '今天什么火' and '热点榜单', which overlap heavily with ordinary conversation and can misroute generic trend questions to this specific Douyin tool. Because the skill queries an external service with a configured token, accidental activation can leak user intent externally and produce platform-specific results the user did not ask for.

Missing User Warnings

Medium

• Confidence: 94% confidence
• Finding: The CLI persistently writes fetched comment data to a local JSON file after completing the request, but the user is not explicitly warned that potentially sensitive or regulated content will be stored on disk. Because this skill is designed for large-scale Douyin comment collection and analysis, the output may include personal data, opinion data, or other sensitive content that can remain accessible to other local users, backups, or downstream tooling.

Missing User Warnings

Medium

• Confidence: 92% confidence
• Finding: The program writes retrieved Douyin results to a local JSON file without any user-facing warning or consent in this file. That creates a privacy and data-handling risk because users may assume the tool only returns data to stdout, while in reality it leaves persistent artifacts that can be accessed by other local users, indexing tools, or backup systems. The skill’s scraping/research use case makes this more concerning because collected datasets may include large volumes of third-party content and identifiers.

Missing User Warnings

Low

• Confidence: 88% confidence
• Finding: The program silently writes search results to local storage without a clear user-facing warning or consent mechanism. In this skill’s context, users may be conducting competitive research, sentiment analysis, or account investigation, so silent persistence can leak sensitive operational interests to other local users, backups, or endpoint monitoring tools.

VirusTotal

VirusTotal findings are pending for this skill version.

Static analysis

No suspicious patterns detected.