← 返回 Skills 市场

Extract Youtube Transcript

Name: Extract Youtube Transcript
Author: hushenglang

作者 Joe Hu · GitHub ↗ · v2.1.0

cross-platform ⚠ suspicious

986

总下载

当前安装

版本数

在 OpenClaw 中安装

/install extract-youtube-transcript

功能描述

Extract plain-text transcripts from YouTube videos using a local Python script. Use when the user wants to fetch, extract, or get a transcript from a YouTube...

使用说明 (SKILL.md)

Extract YouTube Transcript

Fetches plain-text transcripts from YouTube videos using extract_youtube_transcript.py in this skill folder.

Dependency

pip show youtube-transcript-api &>/dev/null || pip install youtube-transcript-api

Quick Start

python extract_youtube_transcript.py "https://www.youtube.com/watch?v=VIDEO_ID"

Supported URL formats: youtube.com/watch?v=, youtu.be/, /embed/, /live/, /shorts/, or a raw 11-char video ID.

Common Patterns

Fetch with preferred language(s)

python extract_youtube_transcript.py "URL" --lang zh-Hant en

Pass languages in priority order. Falls back to any available transcript if none match.

Save transcript to file

python extract_youtube_transcript.py "URL" --output transcript.txt

Text is printed to stdout and also written to the file.

List available languages first

python extract_youtube_transcript.py "URL" --list-langs

Use this to discover what language codes are available before fetching.

Language Codes

Code	Language
`en`	English
`zh-Hant`	Traditional Chinese
`zh-Hans`	Simplified Chinese
`ja`	Japanese
`ko`	Korean
`es`	Spanish

Error Handling

Error	Cause	Recovery
`TranscriptsDisabled`	Owner disabled captions	No transcript available
`NoTranscriptFound`	Requested lang not found	Run `--list-langs`, pick an available code
`VideoUnavailable`	Video is private/deleted	Verify the URL
`AgeRestricted`	Age-gated video	Auth not supported; no workaround
`InvalidVideoId`	Malformed URL or ID	Check the URL format

Workflow

Try a direct fetch first
If NoTranscriptFound, run --list-langs to see available codes, then re-fetch with --lang \x3Ccode>
Save long transcripts to a file with --output for easier downstream processing

安全使用建议

This skill appears to do exactly what it says (locally extract YouTube transcripts), but it ships a Netscape-format cookie file containing long auth tokens. Do NOT use the bundled youtube_cookies.txt as-is: it may contain someone else's session cookies or sensitive tokens. Before using the skill, either remove the youtube_cookies.txt file or replace it with a cookies file you create yourself if you intentionally need authentication (and only if you understand the risks). Review the Python script locally to verify behavior, run it in a sandbox or isolated environment, and avoid uploading any exported transcripts or cookie files to untrusted services. If provenance of this skill is unknown, prefer obtaining the script from a trusted source or recreate minimal functionality rather than running bundled credentials.

功能分析

Type: OpenClaw Skill Name: extract-youtube-transcript Version: 2.1.0 The skill bundle includes a `scripts/youtube_cookies.txt` file containing a full set of active, sensitive YouTube session cookies (including `SID`, `HSID`, `SAPISID`, and `LOGIN_INFO`). While the Python script `extract_youtube_transcript.py` appears to be a legitimate tool for fetching transcripts using the `youtube-transcript-api` and does not contain obvious exfiltration logic, the inclusion of hardcoded, live authentication credentials for a specific Google/YouTube account is a high-risk security anomaly. This could be used for tracking or to facilitate unauthorized access to a specific account, although no direct evidence of intent to harm the user was found.

能力评估

ℹ Purpose & Capability

The name, description, SKILL.md and the Python script are consistent: the tool extracts YouTube transcripts and exposes options to list languages, set output, and optionally supply a cookies file. However, including a pre-filled youtube_cookies.txt (containing many long auth-like cookie values) is not required for the stated purpose and is disproportionate; a transcript extractor should not ship someone else's authentication tokens.

ℹ Instruction Scope

Runtime instructions are confined to installing youtube-transcript-api (via pip) and running the local Python script. They do not instruct broad system reads or external exfiltration. The script supports an optional --cookies argument and will load a Netscape cookie file if provided; the skill bundles such a cookie file which expands the scope to authenticated YouTube access. There is also a pre-scan finding for a 'base64-block' pattern (likely from long cookie values) that should be reviewed.

✓ Install Mechanism

There is no install spec; this is instruction/code-only. Dependency is installed via pip at runtime per SKILL.md. No downloads from arbitrary URLs or archive extraction were found.

⚠ Credentials

The skill declares no required env vars or credentials, which is appropriate. However, it includes a youtube_cookies.txt file containing many cookie tokens (LOGIN_INFO, SID, PSID, etc.) that look like authentication credentials. Bundling those tokens is unnecessary and potentially dangerous — users could mistakenly use someone else's cookies or reveal their own session if they modify the file. The presence of these cookies is disproportionate to the skill's needs and raises privacy/credential risks.

✓ Persistence & Privilege

The skill does not request permanent presence (always:false) and does not attempt to modify other skill or system configurations. It runs locally and does not claim elevated privileges.

如何使用

确保已安装 OpenClaw（本地或 Docker 部署）
在对话框中输入安装命令：/install extract-youtube-transcript
安装完成后，直接呼叫该 Skill 的名称或使用 /extract-youtube-transcript 触发
根据 Skill 的参数说明提供必要输入，即可获得结构化输出

版本历史

v2.1.0

- Bumped version to 2.1.0 in SKILL.md. - No other changes made to documentation or functionality.

v2.0.0

Version 2.0.0 summary: Adds YouTube cookie support for better transcript extraction. - Added support for YouTube cookies to access transcripts that require them (youtube_cookies.txt file). - Updated script and documentation to reflect changes and new usage options.

v1.0.0

Initial release

元数据

Slug extract-youtube-transcript

版本 2.1.0

许可证 —

累计安装 0

当前安装数 0

历史版本数 3

常见问题