Security audit

guaikei-douyin-snapshot-collector

Security checks for vulnerabilities and agentic risk

Overview

This skill mostly matches its Douyin research purpose, but its broad trigger rules could cause unrelated research requests to be sent to and logged through a third-party service.

Review the activation wording before installing in a multi-skill agent. Use this only for clear Douyin or short-video research tasks, and be aware that results are saved under logs/ and that requests go to guaikei.com with the configured API token. The code also emits website/contact guidance on token errors despite SKILL.md saying runtime token errors should avoid contact or promotional content.

SkillSpector

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (6)

Vague Triggers

High

Confidence: 95% confidence
Finding: The description says the skill should trigger even when users only say broad phrases like '查一下' or '帮我看看' without naming Douyin. That creates overbroad activation, increasing the chance the agent invokes this skill in unrelated contexts and unnecessarily sends user queries to an external data-collection workflow.

Vague Triggers

Medium

Confidence: 94% confidence
Finding: The generalized trigger condition states that vague multi-step research tasks should also activate this skill even without mention of Douyin. In a multi-tool agent, this can cause unintended routing of broad research requests to this external collector, leading to unnecessary data exposure, incorrect execution, and user-surprising behavior.

Missing User Warnings

Medium

Confidence: 90% confidence
Finding: The CLI writes the full comment retrieval output, including requested URL and returned comments, to a local JSON file without any visible consent prompt, warning, or opt-in control in this file. Because this skill is specifically designed to collect and inspect social media content at scale, silent persistence increases privacy and data-retention risk and may expose sensitive or regulated content to other local users, backups, or logs.

Missing User Warnings

Medium

Confidence: 90% confidence
Finding: The CLI persists fetched Douyin results to a local JSON file automatically, but this file contains potentially sensitive browsing targets and collected content without any user-facing disclosure or consent prompt. In the context of a snapshot collector, silent local retention increases privacy and data-handling risk because users may not realize search targets and results are being stored on disk for later exposure to other local users, backups, or logs.

Missing User Warnings

Medium

Confidence: 83% confidence
Finding: The tool sends the user-supplied URL and limit to remote API functions to create and fetch tasks, but this file does not explicitly inform the user that their input is transmitted off-host. Because the skill is specifically marketed for researching creators, videos, comments, and trends, the queried targets may reveal user interests or investigative activity, making undisclosed network transmission a meaningful privacy issue.

Missing User Warnings

Low

Confidence: 87% confidence
Finding: The CLI persists full search output to a local JSON file using a filename derived from user-supplied keyword data, without any visible consent or warning in this interface. In the context of a social-media intelligence collection skill, results and query terms can reveal sensitive research interests or collected content, creating privacy and local data exposure risk on shared systems or in downstream log collection.

Static analysis

No suspicious patterns detected.