Security audit

Douyin Comment Reply

Security checks across malware telemetry and agentic risk

Overview

This Douyin auto-reply skill has a coherent purpose, but it needs review because it can publicly reply from a logged-in account and retain commenter data in hard-coded local files.

Install only if you operate the named Douyin account, accept automatic public replies, and are comfortable with commenter data being stored in the specified local Obsidian paths; prefer manual review before posting and verify or change the hard-coded paths before use.

SkillSpector

By NVIDIA

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (3)

Description-Behavior Mismatch

Medium

Confidence: 89% confidence
Finding: The script persists raw comment content, commenter identifiers, video titles, and generated replies into a local Obsidian vault, which creates an undisclosed data-retention path for user-generated content. In this skill context, comments may contain personal data or sensitive messages, so silent local logging expands privacy exposure and increases the blast radius if the workstation or vault is later accessed or synced.

Missing User Warnings

Medium

Confidence: 90% confidence
Finding: The skill stores comment text, usernames, timestamps, account metrics, and operational logs in local JSON and Markdown files, but the description does not warn the user that this data is retained on disk. Because comment content may include personal or sensitive information, silent local retention increases privacy, compliance, and accidental disclosure risk, especially if the Obsidian workspace or temp files are synced, backed up, or shared.

Missing User Warnings

Medium

Confidence: 96% confidence
Finding: The skill is designed to automatically generate and post replies from the user's Douyin account, yet the description does not prominently warn that it will take account actions on the user's behalf. This is more dangerous in context because the skill impersonates a child persona and uses an LLM to decide responses, creating reputational, policy, and safety risks if it posts inappropriate, inaccurate, or noncompliant replies automatically.

VirusTotal

VirusTotal findings are pending for this skill version.

View on VirusTotal

Static analysis

No suspicious patterns detected.