TikTok视频下载

• Outcome: Pass
• Latest audit: Jul 28, 2026 · 4:52 AM UTC
• Version: 0.1.0

Overview

This is a coherent TikTok downloader that uses redfox.hk to process user-provided links, with the main considerations being third-party sharing and optional API-key storage.

Install only if you are comfortable sending TikTok links and your RedFox API key to redfox.hk. Prefer the REDFOX_API_KEY environment variable, keep the key revocable, avoid pasting private or sensitive links, and use small batches because output is not size-limited.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Output Handling: Unvalidated Output Injection, Cross-Context Output, Unbounded Output
• 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

Findings (6)

Missing User Warnings

Medium

• Confidence: 91% confidence
• Finding: The README states that parsing begins immediately once a TikTok link is provided, but it does not clearly disclose that the submitted URL and associated metadata will be sent to a third-party service (RedFoxHub/API backend) for processing. This creates a transparency and consent gap: users may unknowingly transmit browsing interests, creator targets, or research subjects to an external service without explicit notice or opt-in.

Vague Triggers

Medium

• Confidence: 84% confidence
• Finding: The README explicitly tells the agent to act on broad natural-language requests like “下载这个视频 …” and “帮我下载这几个视频 …” without an explicit invocation boundary. In an agent ecosystem, this increases the chance the skill is triggered from ordinary conversation or mixed-content prompts, causing unintended external processing of user-supplied links and surprising actions.

Missing User Warnings

Medium

• Confidence: 96% confidence
• Finding: The skill sends user-supplied TikTok URLs to redfox.hk for parsing, but the description does not clearly warn users that their inputs are transmitted to a third-party service. This creates a privacy and trust risk because users may unknowingly disclose viewing interests, account-linked URLs, or other sensitive metadata to an external provider.

Missing User Warnings

Medium

• Confidence: 92% confidence
• Finding: The workflow explicitly encourages storing an API key in a persistent local config file and does not mention file-permission hardening, secret exposure, or cleanup. This increases the chance that the credential will be recovered from backups, shell history, shared home directories, or accidentally committed files, enabling unauthorized use of the third-party API.

Missing User Warnings

Medium

• Confidence: 89% confidence
• Finding: The skill sends user-supplied TikTok URLs and an API key to an external service but does not clearly warn that data is transmitted off-platform to a third party. Users may unknowingly disclose sensitive URLs, access patterns, or credentials to a remote provider, which is especially relevant in a downloader skill where all functionality depends on external network handling.

Unbounded Output

Medium

• Category: Output Handling

Content

- **Video Parsing**: paste a TikTok video link to parse out the watermark-free video download URL.
- **Batch Parsing**: paste multiple TikTok video links at once (space-separated), parse them one by one and get a summary of successes and failures at the end.
- **Link Validation**: automatically detects whether the input link is a TikTok video link; non-TikTok links are flagged and the process terminates, requiring you to re-enter a valid link.
- **Complete Information**: displays resource type, duration, download link, cover image, audio link, and full description line by line without truncation.
- **Dual Link Support**: recognizes both www.tiktok.com web links and vm.tiktok.com short links.
- **Link Expiry Reminder**: the download link returned after parsing is valid for approximately 5 minutes; users are reminded to copy and use it immediately.

• Confidence: 76% confidence
• Finding: without truncation

VirusTotal

63/63 vendors flagged this skill as clean.

Static analysis

No suspicious patterns detected.