guaikei-douyin-machine-readable-data

• Outcome: Review
• Latest audit: Aug 20, 2026 · 3:34 AM UTC
• Version: 1.0.0

Overview

The skill is mostly a Douyin public-data analysis tool, but it needs Review because it can trigger on generic requests and automatically stores bulk social-media data locally.

Install only if you intentionally want an agent to query Douyin through guaikei.com and save returned JSON locally. Confirm ambiguous market-research requests before using it, avoid collecting more comments or profile data than needed, review/delete the logs directory regularly, and treat returned media URLs and user/comment data as sensitive and subject to platform terms and applicable law.

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 (10)

Description-Behavior Mismatch

Medium

• Confidence: 88% confidence
• Finding: The schema explicitly documents `play_addr` as both a playback and video download address, which conflicts with the stated skill boundary that it does not handle video downloading. This can enable downstream agents or users to use the skill for media retrieval despite policy or product restrictions, increasing risk of unauthorized content acquisition and scope bypass.

Description-Behavior Mismatch

Medium

• Confidence: 94% confidence
• Finding: The CLI writes the full fetched comment dataset to a local JSON file after completing a read-style retrieval operation. Comments can contain personal data, sensitive opinions, usernames, and other scraped content, so persisting them without clear disclosure or opt-in increases privacy and data-retention risk beyond what a user may expect from a simple analysis command.

Description-Behavior Mismatch

Medium

• Confidence: 93% confidence
• Finding: The CLI writes the full search output, including user-provided keywords and returned results, to a local file without any visible consent, retention policy, or minimization. In a data-collection skill that may process sensitive research topics, competitor targets, or comment-derived insights, this creates an unintended persistence channel and increases exposure to local compromise, overcollection, and accidental disclosure.

Vague Triggers

Medium

• Confidence: 88% confidence
• Finding: The description says the skill should be used for requests like competitor analysis or 'what's hot in the industry' even when Douyin is not mentioned. That overly broad trigger scope can cause the agent to invoke this skill for generic research tasks, leading to unintended collection of Douyin data, user-surprising behavior, and possible misuse of the tokenized external service.

Vague Triggers

Medium

• Confidence: 86% confidence
• Finding: The implicit trigger section includes generic phrases such as short-video competitor research, high-performing content discovery, and 'what's popular lately,' which overlap with many ordinary research requests unrelated to Douyin. This increases the chance of inappropriate auto-invocation, causing context drift, unnecessary external queries, and data access beyond user intent.

Vague Triggers

Medium

• Confidence: 92% confidence
• Finding: The skill maps broad everyday phrases such as '帮我做个竞品分析' and '看看行业什么火' to Douyin data collection even when the user does not mention Douyin. This can cause overbroad or unintended activation, leading the agent to access third-party data sources and collect/export social media data when the user may have meant a generic analysis request.

Vague Triggers

Medium

• Confidence: 95% confidence
• Finding: The intent recognition rules rely on generic keywords such as '热搜', '搜索', '评论', '作品', and '账号' without enough scoping constraints. In an agent environment, this increases the chance of false activation and unintended scraping of public author, post, or comment data, especially because these terms are common in many non-Douyin contexts.

Missing User Warnings

Medium

• Confidence: 90% confidence
• Finding: The readme promotes bulk collection and export of public video, author, comment, and hot-list data, but it does not clearly warn about privacy, retention, lawful use, or handling of personal data in comments and profiles. Because the skill supports large-scale extraction and JSON export, users may unknowingly process personal data in ways that create privacy, compliance, or misuse risks.

Missing User Warnings

Medium

• Confidence: 91% confidence
• Finding: This file silently stores retrieved comment results to disk without an explicit user-facing warning in the command flow. In a data-collection skill focused on analyzing Douyin content, that behavior is more dangerous because comment text may include user-generated personal or sensitive content, and silent persistence can expose it to other local users, backups, or later misuse.

Missing User Warnings

Medium

• Confidence: 95% confidence
• Finding: Search results are persisted locally without a user-facing warning, even though the tool presents itself as a search/analysis utility rather than a data-retention mechanism. Because results may contain sensitive business research, account targets, or user-generated content, silent file persistence can surprise users and create privacy, compliance, and workstation data-leak risks.

VirusTotal

64/64 vendors flagged this skill as clean.

Static analysis

No suspicious patterns detected.