A skill that automates repurposing Chinese social videos (Douyin/Bilibili/Xiaohongshu) to international platforms (TikTok/YouTube/Instagram) via the Lumi API — handling translation, AI dubbing, and publishing in one workflow.

• Outcome: Pass
• Latest audit: May 29, 2026 · 8:25 PM UTC
• Version: 1.0.0

Overview

This is a disclosed Lumi social-media automation skill that can upload, translate, and publish videos, including automatic publishing after the user selects target platforms.

Install only if you intend to let Lumi manage and potentially publish videos through your connected social accounts. Use a dedicated Lumi API key, verify the selected account, platform, caption/title, and visibility before posting, and use the localization-only path when you want to review the processed video before publication.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Trigger Abuse: Overly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access
• Supply Chain: Unpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (5)

Vague Triggers

Low

• Confidence: 86% confidence
• Finding: The API operation enables broad external publishing to multiple social platforms but the spec provides no invocation guardrails, approval requirements, or contextual constraints for when an agent should call it. In an agent setting, this increases the risk of unintended or prompt-induced posting to public third-party accounts, which can cause reputational damage and unauthorized content publication.

Missing User Warnings

Medium

• Confidence: 93% confidence
• Finding: This endpoint publishes content to external social networks, an action that may be public, difficult to reverse, and capable of affecting brand reputation or compliance posture, yet the spec does not require any user-facing warning or confirmation. In the context of an agent skill, missing disclosure makes accidental, misunderstood, or socially engineered posting materially more dangerous because the consequence extends beyond the local system to real public accounts.

Vague Triggers

Medium

• Confidence: 88% confidence
• Finding: The upload endpoint description is broad enough that an agent may invoke it on any user-provided public URL without clearly confirming that the remote content will be fetched, copied into Lumi-managed storage, and retained in the user's library. In an agent skill context, ambiguous action semantics can cause unintended third-party data ingestion or copyright/privacy violations because the user may believe they are only referencing a URL rather than initiating transfer and storage.

Missing User Warnings

Medium

• Confidence: 93% confidence
• Finding: There is no user-facing warning that submitting a URL triggers server-side fetching and persistence to OSS and the video library. In this skill's context, that increases the chance of covert exfiltration of private-but-publicly-accessible media, accidental processing of copyrighted content, or user surprise about storage and downstream reuse in publishing/localization workflows.

Missing User Warnings

Medium

• Confidence: 95% confidence
• Finding: The Chinese-platform endpoint explicitly crawls third-party platforms to extract real video URLs and uploads the resulting media to OSS, yet the spec does not surface a warning to the user about that behavior. This is more dangerous in a social-media automation skill because an agent could silently copy content from Douyin/Bilibili/Xiaohongshu into managed storage, creating legal, privacy, and consent risks beyond a simple link-handling operation.

VirusTotal

64/64 vendors flagged this skill as clean.

Static analysis

No suspicious patterns detected.