Security audit

guaikei-douyin-collect-public-data

Security checks across malware telemetry and agentic risk

Overview

This skill is a mostly coherent Douyin public-data collector, but it can trigger too broadly and automatically leaves collected social data in local log files.

Install only if you intend to collect Douyin public data through guaikei.com. Treat searches, video URLs, creator IDs, comments, user nicknames/IDs, and IP-region labels as data that may be sent to the API provider and stored locally under the skill's logs folder; delete or protect those logs as needed and avoid using the skill for tasks that did not explicitly call for Douyin research.

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 (5)

Description-Behavior Mismatch

Medium

Confidence: 88% confidence
Finding: The CLI persists retrieved comment data to a local JSON file after printing the results, creating an undeclared secondary storage path for potentially sensitive user-generated content. In a public-data collection skill, silent persistence increases data-retention and privacy risk, especially if comments contain personal information or if the host environment is shared.

Vague Triggers

High

Confidence: 94% confidence
Finding: The manifest says the skill should trigger even when the user's task does not explicitly mention Douyin, which broadens activation beyond clear user intent. That can cause the agent to invoke an external data-collection/API skill unexpectedly, leading to unnecessary data disclosure to a third-party service, irrelevant actions, or confused task routing.

Missing User Warnings

Medium

Confidence: 84% confidence
Finding: The README explicitly states that collected Douyin data is automatically exported to local JSON log files, but it does not present a prominent privacy and data-handling warning before use. Because the skill collects public videos, authors, comments, and trend data at scale, stored logs may contain personal data or sensitive profiling information that can be retained, shared, or accessed unintentionally.

Missing User Warnings

Medium

Confidence: 90% confidence
Finding: The tool writes fetched comment results to disk without an explicit user warning or consent, which can surprise users and leave collected data behind on the local system. Because the skill is framed as reading/analyzing public comments, undisclosed persistence makes the behavior more dangerous by expanding data exposure beyond the immediate command output.

Missing User Warnings

Medium

Confidence: 92% confidence
Finding: The CLI writes the full result set to a local JSON file after fetching Douyin post data, but there is no explicit user consent, warning, or control over persistence. Even though the skill targets public data, the output can still include large-scale collected metadata and derived intelligence that may be sensitive in operational, research, or shared-host environments, and local files can be exposed to other users, backups, or malware.

VirusTotal

61/61 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.