Security audit

douyin-creator-insight

Security checks across malware telemetry and agentic risk

Overview

The skill appears purpose-built for public Douyin creator reports, but it should be reviewed because it uses an authenticated persistent browser profile and preserves broad raw data.

Install only if you are comfortable letting the skill use an existing Douyin browser session to read public creator pages. Prefer a separate low-privilege browser profile, do not provide personal cookies to third-party actors, keep generated JSON reports private, and delete raw outputs when no longer needed.

SkillSpector

By NVIDIA

Vulnerability Patterns

Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access

Findings (9)

Context-Inappropriate Capability

Medium

Confidence: 92% confidence
Finding: The playbook suggests using B站/YouTube as substitute sources when Douyin transcripts are unavailable, which expands the skill beyond its declared Douyin-focused scope. This can cause data-mixing, inaccurate attribution, and unintended external collection behavior that users may not expect from this skill.

Description-Behavior Mismatch

Medium

Confidence: 91% confidence
Finding: The code explicitly reuses a persistent browser profile from a different skill, one intended for favorites syncing, which likely contains authenticated cookies and broader account state. In a skill advertised as analyzing public creator data, coupling to another skill’s logged-in profile expands access beyond the declared scope and creates a risk of unintended access to private/session-bound data or cross-skill data leakage.

Context-Inappropriate Capability

Medium

Confidence: 95% confidence
Finding: The collector relies on an authenticated persistent browser session to fetch data, even though the stated purpose is collecting public creator information. This mismatch is dangerous because it encourages use of logged-in credentials where unauthenticated access should suffice, increasing the chance of overcollection, account-linked requests, and accidental access to data unavailable to the public.

Missing User Warnings

Medium

Confidence: 91% confidence
Finding: The document notes that some fallback actors may require login cookies, but it does not explicitly warn users not to provide personal/session cookies casually, nor does it describe the account-takeover and data-exposure risks of sharing authenticated credentials with third-party scraping services. In a scraping workflow, this omission can lead operators to paste live session cookies into external actors, exposing private account access and potentially violating platform policies.

Missing User Warnings

Medium

Confidence: 94% confidence
Finding: Telling users to upload raw JSON/manual data without warning about sensitive fields can lead to collection of more data than necessary, including account identifiers, URLs, metadata, or other personal information embedded in exported datasets. The skill context makes this more relevant because creator-analysis workflows often retain raw snapshots for debugging and reprocessing.

Missing User Warnings

Medium

Confidence: 95% confidence
Finding: The instruction to preserve and reuse raw JSON encourages retention of full upstream datasets without any notice about sensitivity, retention duration, or minimization. Retaining raw data increases exposure if logs, artifacts, or cached files are later accessed by unauthorized parties or reused outside the original purpose.

Missing User Warnings

Low

Confidence: 83% confidence
Finding: Suggesting Bilibili/YouTube as alternative inputs without warning broadens data collection scope and may surprise users who expect Douyin-only analysis. While the data is still described as public, the lack of disclosure weakens consent and provenance clarity.

Missing User Warnings

Medium

Confidence: 96% confidence
Finding: The offline-mode advice to upload historical raw JSON or authorized snapshots omits basic handling expectations for potentially sensitive content, despite encouraging transfer and processing of stored datasets. This creates unnecessary privacy risk through over-collection, unclear authorization assumptions, and indefinite reuse of artifacts.

Missing User Warnings

Medium

Confidence: 92% confidence
Finding: The function transmits Douyin video URLs to third-party Apify actors with no visible consent, disclosure, or trust boundary enforcement in this file. Even if the target videos are public, this creates a data-sharing/privacy risk and expands exposure to external services whose retention, logging, or downstream processing may not match user expectations.

VirusTotal

61/61 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.