TikTok/Douyin 创作流水线

TikHub API 多平台数据爬取工具，支持抖音/TikTok/B站等。当用户提到：(1) 爬取抖音/TikTok/B站视频或评论；(2) 获取用户信息/粉丝列表；(3) 批量下载无水印视频；(4) 抖音链接转文字（下载→音频→Whisper pipeline）；(5) 调用 TikHub API。

This skill appears to implement what it claims (TikHub scraping, downloading, audio extraction, Whisper/MLX-Whisper transcription). Before installing or running it: - Review how you supply the TikHub API key — avoid hardcoding keys in the file; prefer passing via set_api_key at runtime or environment variables under your control. - The code uses subprocess with shell=True to launch background Whisper jobs (nohup). If you pass filenames or URLs containing untrusted characters, that could allow command injection. Only run the skill on files/inputs you trust or sanitize paths before use. - The skill spawns background processes and writes logs to /tmp; monitor for orphaned processes or large disk usage. - requirements.txt incorrectly includes 'ffmpeg' (not a pip package). Follow SKILL.md (brew apt-get or OS package manager) to install ffmpeg and the Python dependencies (requests, openai-whisper, mlx-whisper). - Consider inspecting the full set_api_key implementation (in scripts/tikhub.py) to confirm it does not persist secrets insecurely. - Legal/terms note: scraping/downloading videos and comments may violate service terms or local laws — ensure you have permission and are within terms of service. If you need to proceed but want to reduce risk: run the code in an isolated environment (container/VM), inspect or patch the subprocess usage to avoid shell=True (use list-argument subprocess.run without shell), and validate/sanitize all filenames and URLs before passing them to shell commands.

Type: OpenClaw Skill Name: tikhub-douyin-pipeline Version: 1.1.0 The skill bundle contains a critical shell injection vulnerability in `scripts/tikhub.py` within the `whisper_transcribe` function. This function constructs a command string using unsanitized input (an `aweme_id` derived from a user-provided URL) and executes it via `subprocess.run(..., shell=True)`. An attacker could exploit this to achieve Remote Code Execution (RCE) by providing a crafted URL containing shell metacharacters. While the code appears to be a legitimate tool for the TikHub API (interacting with `api.tikhub.io` and `api.tikhub.dev`), the lack of input sanitization in a shell-executing context is a major security flaw.