bilibili-up-batch

Batch-download recent Bilibili videos from a UP page and generate per-video analysis evidence.

Updated Dec 17, 2025
One-click install
npx skills add https://github.com/haizhouyuan/codexreadpublic --skill bilibili-up-batch
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: bilibili-up-batch
Source: https://github.com/haizhouyuan/codexreadpublic/tree/main/skills-src/bilibili-up-batch
Command: npx skills add https://github.com/haizhouyuan/codexreadpublic --skill bilibili-up-batch

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) components.

What problem does it solve?

This Skill batch-downloads recent Bilibili videos, runs video analysis, and prepares evidence packets for downstream digesting.

Core Features & Use Cases

  • Batch download: Fetch latest videos for a given UP, store locally.
  • Evidence generation: Run video_pipeline to create evidence.json and evidence_compact.md per video.
  • Digest readiness: Optionally feed results to digest-content for rapid digest creation.

Quick Start

python3 scripts/bilibili_up_batch.py --up '<space_or_b23_url>' --limit 30 --download --analyze --enable-ocr

Frequently Asked Questions about bilibili-up-batch

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I batch download Bilibili videos for analysis?

Batch download the latest N videos from a Bilibili UP page using the script with a space.bilibili.com or b23.tv URL, specifying --limit and --download flags. Videos store locally under imports/content/videos/bilibili/<mid>/ for subsequent analysis without cluttering git.

What do I need installed to run video analysis on Bilibili downloads?

Install ffmpeg and yt-dlp before running batch downloads. These tools handle video retrieval and processing; the Skill operates on local downloads only and requires both utilities present in your environment.

Can I generate evidence summaries from downloaded Bilibili videos?

Yes, the --analyze flag runs video_pipeline on each downloaded video to generate evidence.json and evidence_compact.md per video. Enable --enable-ocr for optical character recognition within the analysis output.

What happens to the analysis output after batch processing completes?

Analysis results store under state/video-analyses/<analysis_id>/evidence.json and evidence_compact.md. You can optionally feed these outputs to digest-content for rapid digest creation and archiving.

Do I need to commit downloaded videos to version control?

No, the Skill operates on local downloads only—video files are not git-tracked. Store outputs under imports/content/videos/bilibili/<mid>/ to keep your repository clean.

Can I analyze multiple UP pages in a single run?

The script processes one UP page URL per invocation. Run the command multiple times with different space.bilibili.com/<mid> or b23.tv URLs to batch-process multiple creators sequentially.