guaikei-douyin-extract-videos-comments-hot

• Outcome: Review
• Latest audit: Aug 18, 2026 · 2:13 AM UTC
• Version: 1.0.0

Overview

The skill mostly does what it claims, but it can automatically send broad short-video research requests to a third-party API and saves fetched public user/comment data to local logs despite inconsistent disclosure.

Review before installing. Use this only for intentional Douyin research, keep the API token private, and treat generated logs as retained data that may contain personal information from public profiles or comments. Delete logs when no longer needed and avoid using the skill for generic short-video requests unless Douyin is actually the intended platform.

SkillSpector

By NVIDIA

Vulnerability Patterns

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

Findings (10)

Description-Behavior Mismatch

Medium

• Confidence: 91% confidence
• Finding: The schema explicitly documents `play_addr` as both a playback URL and a video download address, which conflicts with the skill description stating it does not support downloading videos. This mismatch can enable downstream agents or integrators to expose or automate downloading behavior that the skill policy intended to prohibit, creating policy bypass and misuse risk.

Description-Behavior Mismatch

Low

• Confidence: 86% confidence
• Finding: The CLI persists fetched comment data to a local JSON file after already printing the results, creating an additional undisclosed storage location for potentially sensitive or regulated user-generated content. In the context of a public-data retrieval skill, this increases data retention and exposure risk because comments may contain personal information, and local files can be accessed by other users, processes, backups, or later tooling.

Description-Behavior Mismatch

Medium

• Confidence: 93% confidence
• Finding: The CLI writes the full fetched results object to a local file via log.taskWrite, even though the skill is described primarily as public-data extraction. Persisting harvested data creates an additional data-retention and disclosure surface: scraped content, metadata, and potentially user-provided identifiers may remain on disk longer than expected, be readable by other local users/processes, or be unintentionally collected by backups and telemetry.

Vague Triggers

Medium

• Confidence: 92% confidence
• Finding: The skill instructs activation even when the user does not explicitly mention Douyin, based only on generic short-video, blogger, comment, or hot-topic requests. That broad trigger can cause the agent to invoke a platform-specific data collection skill without confirming user intent, leading to incorrect tool use, unintended third-party API calls, and disclosure of user prompts or targets to an external service.

Vague Triggers

Medium

• Confidence: 94% confidence
• Finding: The example workflow operationalizes the same broad auto-triggering by telling the agent to use Douyin workflows for a generic short-video topic request ('美妆赛道') even when the user never specified the platform. This increases the likelihood of systematic misrouting to an external API and unnecessary processing of user research targets by a third party, making the overbroad trigger materially more dangerous in practice.

Missing User Warnings

Medium

• Confidence: 94% confidence
• Finding: The README promotes bulk extraction of public videos, authors, comments, and hot-list data, but does not clearly warn that the returned data and auto-saved JSON logs can contain personal data, opinion data, usernames, profile identifiers, and other sensitive content. In a scraping/monitoring skill focused on comments and competitor analysis, this omission increases the risk of accidental over-collection, insecure retention, and downstream misuse of personal information.

Missing User Warnings

Low

• Confidence: 83% confidence
• Finding: The README tells users to place the API token in an environment variable, but does not prominently warn against exposing the token through shell history, screenshots, shared terminals, CI logs, or committing secrets into config files. While environment variables are a common mechanism, the absence of explicit credential-handling guidance can lead to token leakage and unauthorized API use.

Missing User Warnings

Medium

• Confidence: 82% confidence
• Finding: The documentation explicitly enables bulk collection of Douyin creator works and public comments, but provides no guidance on lawful basis, platform Terms of Service, retention limits, or handling of personal data in comments/user profiles. In a scraping-focused skill, that omission increases the likelihood of privacy misuse, over-collection, and downstream noncompliant processing of user-generated content.

Missing User Warnings

Medium

• Confidence: 91% confidence
• Finding: The code silently writes retrieved comments to a local file without a clear user-facing warning or consent mechanism. Because comment content can include personal data, opinions, usernames, or other sensitive text, undisclosed persistence materially increases privacy and compliance risk beyond transient retrieval.

Missing User Warnings

Medium

• Confidence: 94% confidence
• Finding: The CLI writes the full search output to a local JSON file whose filename includes the user-supplied keyword, but the user is not clearly warned that query terms and retrieved results will be persisted on disk. In this skill context, search terms and returned public-platform data may still contain sensitive research topics, account targets, or personal data from comments/content analysis, creating unintended local data retention and possible exposure to other local users, backups, or log collectors.

VirusTotal

VirusTotal findings are pending for this skill version.

Static analysis

No suspicious patterns detected.