guaikei-douyin-public-content-to-dataset-export

• Outcome: Review
• Latest audit: Aug 21, 2026 · 3:22 AM UTC
• Version: 1.0.0

Overview

The skill broadly collects Douyin public data through a token-backed API and saves datasets locally, with some scope and disclosure mismatches users should review before installing.

Install only if you intend agents to use a GUAIKEI token to query Douyin public data and save JSON datasets locally. Treat saved logs as containing potentially personal public metadata, delete them when no longer needed, and confirm Douyin use when a request is ambiguous. Do not use the returned media URL fields for video downloading or redistribution.

SkillSpector

Vulnerability Patterns

• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Trigger Abuse: Overly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
• MCP Least Privilege: Underdeclared Capability, Wildcard Permission, Missing Permission Declaration
• MCP Tool Poisoning: Hidden Instructions, Unicode Deception, Parameter Description Injection
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands

Findings (11)

Lp3

Medium

• Category: MCP Least Privilege
• Confidence: 83% confidence
• Finding: The skill explicitly depends on an environment variable (`GUAIKEI_API_TOKEN`) but does not declare corresponding permission requirements in a machine-readable way. This can cause the agent platform or reviewer to underestimate the skill’s access to secrets, weakening transparency and least-privilege controls around token handling.

Description-Behavior Mismatch

Medium

• Confidence: 92% confidence
• Finding: The schema explicitly describes `video_url` as both playback and download addresses, which conflicts with the stated skill boundary that it is not for downloading videos. Even without executable code, documenting and normalizing download URLs into the API output makes downstream misuse easier and can enable users or integrators to repurpose the skill for prohibited content acquisition.

Description-Behavior Mismatch

Medium

• Confidence: 90% confidence
• Finding: The schema explicitly documents `play_addr` as both a playback address and a video download address, which conflicts with the skill's stated restriction that it is not for downloading videos. Exposing or normalizing a direct download field can enable downstream agents or users to retrieve video files anyway, weakening policy boundaries and increasing the risk of misuse of copyrighted or platform-restricted content.

Description-Behavior Mismatch

Medium

• Confidence: 91% confidence
• Finding: The CLI writes full search results to a local JSON file, creating data persistence beyond the immediate query operation. Even if the content is publicly sourced, the saved dataset can include user search intent and large-scale collected records, which increases privacy, retention, and local data exposure risk if the host is shared or logs are later exfiltrated.

Vague Triggers

High

• Confidence: 95% confidence
• Finding: The description says the skill should be used even when the user does not mention Douyin, creating an overly broad trigger surface. That can cause the agent to invoke a data-collection skill in contexts where the user intended general research, increasing the chance of unintended third-party queries, unnecessary token use, and data access beyond user expectation.

Vague Triggers

High

• Confidence: 98% confidence
• Finding: The generalized trigger section explicitly instructs activation without explicit platform mention, which is a strong prompt-routing risk. In context, this skill performs external public-data collection and uses an API token, so misrouting is more dangerous than for a purely local formatting skill because it can cause unintended external actions and secret-backed API consumption.

Missing User Warnings

Medium

• Confidence: 90% confidence
• Finding: The README explicitly promotes bulk collection and JSON export of public videos, author data, comments, and hot-list data, but it omits privacy/compliance handling guidance. Even when data is publicly accessible, exported comments and account metadata can contain personal information; encouraging local dataset creation without minimization, retention, or lawful-use warnings increases risk of privacy misuse and non-compliant downstream sharing.

Missing User Warnings

Low

• Confidence: 80% confidence
• Finding: The README tells users to set GUAIKEI_API_TOKEN as an environment variable but does not warn against exposing the token via shell history, screenshots, process output, CI logs, or checked-in config files. This can lead to credential leakage and unauthorized API use, especially because the skill is aimed at easy CLI execution and log generation.

Vague Triggers

Medium

• Confidence: 76% confidence
• Finding: The natural-language activation examples are broad enough to match common user phrases such as generic requests to search or find videos, which can cause over-broad or unintended invocation of a scraping/export skill. In this skill's context, accidental activation is more sensitive because the tool performs third-party data collection and local dataset export, potentially initiating collection users did not specifically intend.

Missing User Warnings

Medium

• Confidence: 88% confidence
• Finding: The CLI writes the fetched Douyin results to a local JSON file automatically after successful execution, without any visible opt-in, warning, or control in this file. Because the skill is specifically designed to collect public-content datasets, the stored output may contain usernames, profile identifiers, comments, or other scraped metadata that can persist on disk unexpectedly and be exposed to other local users, backup systems, or downstream tooling.

Missing User Warnings

Medium

• Confidence: 90% confidence
• Finding: Results are saved locally without an explicit user-facing disclosure in the CLI flow, which can surprise users who expect a transient search tool. Silent persistence can lead to unintended collection of search-derived datasets and accidental exposure of saved content or user activity on disk.

VirusTotal

VirusTotal findings are pending for this skill version.

Static analysis

No suspicious patterns detected.