guaikei-douyin-deep-comments

• Outcome: Review
• Latest audit: Aug 20, 2026 · 2:34 AM UTC
• Version: 1.0.0

Overview

This skill is mostly a Douyin public-data research tool, but it needs Review because it can trigger on generic research requests, sends queries to a third-party API, and saves scraped data locally by default.

Install only if you are comfortable sending Douyin search terms, target URLs, and usage metadata to guaikei.com using your token. Treat GUAIKEI_API_TOKEN as a secret, avoid generic auto-invocation for unspecified research tasks, and regularly review or delete the generated logs directory because it may contain scraped public comments, profile metadata, and sensitive search interests.

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: 91% confidence
• Finding: The skill explicitly depends on an environment variable (`GUAIKEI_API_TOKEN`) and invokes Node.js commands, but no declared permissions are shown in the skill metadata. This creates a transparency and policy-enforcement gap: reviewers and runtime controls may not realize the skill needs access to secrets, increasing the risk of unintended secret exposure or over-broad execution in a host environment.

Description-Behavior Mismatch

Medium

• Confidence: 84% confidence
• Finding: The schema explicitly documents `play_addr` as both a playback and download URL, which expands the skill’s capability beyond its stated scope of not supporting download-oriented use. In a scraping/research skill for Douyin, exposing a direct media retrieval field can enable unauthorized downloading, policy circumvention, or downstream misuse even if the manifest claims otherwise.

Description-Behavior Mismatch

Medium

• Confidence: 93% confidence
• Finding: The CLI persists fetched comment data to a local JSON file after completing a nominally read-oriented retrieval action. Comment data can contain personal or sensitive content, and writing it to disk creates an additional storage surface, retention risk, and possible unintended disclosure if the host environment is shared or logs/artifacts are collected automatically.

Vague Triggers

Medium

• Confidence: 87% confidence
• Finding: The description says the skill should also apply when users are doing 'content research' without naming the platform, which is an overly broad implicit trigger. This can cause the agent to invoke the Douyin data-collection skill in unrelated research contexts, potentially sending user queries or URLs to the wrong tool and collecting external data the user did not intend to target.

Vague Triggers

Medium

• Confidence: 90% confidence
• Finding: The implicit trigger examples include generic phrases such as '帮我做短视频竞品调研' and '最近什么热门', which are common across many platforms and tasks. In an agentic environment, this raises the chance of accidental tool activation and unintended data access/workflow execution, especially because the skill can query external services and process URLs/comments.

Missing User Warnings

Medium

• Confidence: 91% confidence
• Finding: The README explicitly promotes automatic JSON log export of scraped Douyin data but does not warn that the exported files may contain large amounts of public user-generated content, profile metadata, or comment text that can become sensitive when aggregated and retained locally. In a scraping-oriented skill, silent persistence of collected data increases the risk of unintended disclosure, over-retention, and misuse of harvested content.

Missing User Warnings

Medium

• Confidence: 94% confidence
• Finding: The README instructs users to place GUAIKEI_API_TOKEN in an environment variable but does not clearly foreground that the token is a sensitive credential that must not be logged, hardcoded, shared, or committed to source control. Because this skill relies on a third-party API token and is aimed at easy CLI use, weak credential-handling guidance can lead to accidental leakage and unauthorized use of the service.

Missing User Warnings

Medium

• Confidence: 84% confidence
• Finding: The tool sends user-supplied Douyin URLs and query parameters to an external service using an API token, but the CLI flow does not provide a clear user-facing notice or explicit consent boundary for external transmission. In a data-research context, this can expose queried targets, usage patterns, or potentially sensitive investigative inputs to a third-party backend.

Missing User Warnings

Medium

• Confidence: 72% confidence
• Finding: The tool silently writes fetched Douyin post data to a local JSON file without explicit user consent or a prominent disclosure at runtime. In a research or sentiment-analysis context, the returned data may include account content, metadata, or other collected records that become unexpectedly persistent on disk, increasing the risk of leakage through shared machines, backups, or later unauthorized access.

Missing User Warnings

Medium

• Confidence: 94% confidence
• Finding: The CLI persists full search requests and returned Douyin content to a local JSON file named with the user keyword, but this file gives no warning or consent flow about local retention. That creates a privacy and data-handling risk because search terms may be sensitive and filenames plus stored results can expose user interests or collected third-party content to other local users, backups, or downstream tooling.

Missing User Warnings

Medium

• Confidence: 86% confidence
• Finding: The script transmits user-supplied keywords and search parameters to a remote API to perform the requested search, but this file does not disclose that network transmission at runtime. In a data-collection skill, that can still be risky because users may input sensitive research terms, and those terms are then shared with an external service without explicit notice in the command flow.

VirusTotal

64/64 vendors flagged this skill as clean.

Static analysis

No suspicious patterns detected.