Security audit

guaikei-douyin-public-data-for-market-analysis

Security checks across malware telemetry and agentic risk

Overview

This Douyin analysis skill mostly matches its purpose, but needs review because it can trigger on broad research prompts, sends inputs to guaikei.com, and saves full result logs by default.

Review before installing. Use it only when you intend to query Douyin public data through guaikei.com, avoid broad/ambiguous prompts unless you want Douyin data, and periodically delete or protect the logs folder because it can contain your research terms, target accounts, comments, and returned public-content metadata.

SkillSpector

By NVIDIA

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access

Findings (7)

Description-Behavior Mismatch

Medium

Confidence: 86% confidence
Finding: The CLI writes the full scraped results to a local JSON file using a filename derived from the target account URL. Even though the skill is for public-data analysis, local persistence creates an undeclared data-retention surface: scraped content may include large datasets, usernames, captions, and other metadata that remain on disk after execution and can be accessed by other local users, backup systems, or later processes.

Vague Triggers

Medium

Confidence: 88% confidence
Finding: The generalized trigger text says the skill should activate for broad research tasks even when the user did not mention Douyin, which can cause overbroad invocation and unnecessary transfer of user queries to an external data service. In a multi-skill agent, this increases the risk of scope creep, unintended data disclosure, and user confusion about which platform is being queried.

Missing User Warnings

Medium

Confidence: 93% confidence
Finding: The skill states that outputs are automatically saved under `logs/` with timestamps and query parameters, but it does not prominently warn users that their search terms, URLs, and resulting data will be written to disk. This can expose potentially sensitive business research, target account lists, or user-provided URLs to other local users, backups, or later forensic access.

Vague Triggers

Medium

Confidence: 88% confidence
Finding: The hot-search trigger mapping uses very broad everyday phrases such as '热点' and '今天什么火', which can match normal conversation and cause the skill to run when the user did not clearly request Douyin data retrieval. In an agent setting, this creates unintended tool invocation risk and may trigger external API calls, data collection, and logging without sufficiently specific user intent.

Vague Triggers

Medium

Confidence: 84% confidence
Finding: The search trigger definitions include colloquial phrases like '搜索 / 搜一下 / 找 + 关键词' that are overly generic for a skill integrated into a broader agent. Such broad matching can cause accidental invocation from ordinary user requests, leading to unnecessary external queries and storage of results in local logs.

Missing User Warnings

Medium

Confidence: 80% confidence
Finding: The README advertises automatic JSON log and output generation but does not prominently warn that query terms, retrieved public-content data, and potentially sensitive analysis targets will be written to disk. In practice, this can expose research interests, monitored accounts, and collected comments to other local users, backups, or downstream systems without the operator fully realizing it.

Missing User Warnings

Medium

Confidence: 78% confidence
Finding: The CLI sends the user-supplied Douyin URL to a remote API (`createPostTask` / `getPostTask`) and later persists returned results locally, but this file provides no explicit notice or consent mechanism for either network transmission or disk storage. In a data-collection tool this is a transparency and privacy/security issue because users may assume the tool only processes input locally or may not expect durable local logs.

VirusTotal

VirusTotal findings are pending for this skill version.

View on VirusTotal

Static analysis

No suspicious patterns detected.