agentleFS
chengkj99chengkj99/kj-skills/skills/kangjian-douyin-media/SKILL.md

kangjian-douyin-media

Skill14 starsChanged 5 months agoMITView on GitHub

Incrementally crawl, archive, and analyze 程序员康健 / kangjian-douyin Douyin creator media data with MediaCrawler. Use when the user asks to update 程序员康健's Douyin account, crawl new Douyin videos, refresh published Douyin works/comments, generate Douyin full-ledger or review-analysis Markdown reports, or maintain raw/studio/funnel/published/douyin snapshots.

The file

Show its rules and text

We split this skill into the rules it gives an agent

17 rules. About 0 are common, 5 are distinctive, and 12 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈5

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 12

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Use this skill to update the user's own Douyin archive for:

    line 8
  2. Very specific:

    Default archive root in the wiki: raw/studio/funnel/published/douyin/

    line 13
  3. Very specific:

    Use NanmiCoder/MediaCrawler to produce raw Douyin JSONL for works and comments.

    line 17
  4. Very specific:

    Run scripts/update_douyin_archive.py to merge the run incrementally and generate Markdown reports.

    line 18
  5. Very specific:

    Do not download media files by default.

    line 20
  6. Very specific:

    Use an existing MediaCrawler checkout if available.

    line 32
  7. Very specific:

    Use --crawler_max_notes_count 130 or higher for a full-account refresh.

    line 51
  8. Distinctive:

    After MediaCrawler finishes, run:

    line 55
  9. Very specific:

    read douyin/jsonl/creator_comments_*.jsonl when present

    line 67
  10. Very specific:

    create or refresh latest-kangjian-douyin-summary.md

    line 71
  11. Very specific:

    Use --full-refresh when the run intentionally covers the whole account.

    line 73
  12. Very specific:

    Run wc -l on contents and comments JSONL.

    line 98
  13. Very specific:

    Run the archive update script and report its works, comments, and new_works counts.

    line 99
  14. Very specific:

    If MediaCrawler fails from missing dependencies, run uv sync in the checkout and retry.

    line 105
  15. Distinctive:

    If browser or login fails, ask the user to approve Chrome/QR login again.

    line 106
  16. Distinctive:

    If comments are incomplete due to rate limiting, keep the run but say it is partial and rerun later with lower concurrency.

    line 107
  17. Distinctive:

    If the archive path is outside the current writable workspace, request filesystem approval instead of copying data silently.

    line 108

≈ “Common” is an estimate: at least 1 in 5 of the 372,429 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Kangjian Douyin Media

Use this skill to update the user's own Douyin archive for:

  • Account: kangjian-douyin
  • Display identity: 程序员康健
  • Homepage: https://www.douyin.com/user/MS4wLjABAAAAU0QYHCmBaS2uUvbuc5XfY-XTJJCybRun-GHf2kpR6o0?from_tab_name=main
  • Default archive root in the wiki: raw/studio/funnel/published/douyin/

The skill has two layers:

  1. Use NanmiCoder/MediaCrawler to produce raw Douyin JSONL for works and comments.
  2. Run scripts/update_douyin_archive.py to merge the run incrementally and generate Markdown reports.

Do not download media files by default. Do not fetch sub-comments by default. Transcripts are produced separately with $local-stt-transcription and should be linked under raw/studio/funnel/transcripts/.

Update Cadence

  • New video: crawl after 24 hours for first feedback.
  • Review window: crawl again after 3-7 days.
  • Baseline refresh: crawl the full account monthly or quarterly.

For normal new-video updates, crawl the latest 30-50 works. For audits or drift correction, crawl the full account with a larger cap.

MediaCrawler Run

Use an existing MediaCrawler checkout if available; otherwise clone it to /private/tmp/MediaCrawler-kangjian-douyin.

Recommended command shape:

uv run main.py \
  --platform dy \
  --lt qrcode \
  --type creator \
  --creator_id "https://www.douyin.com/user/MS4wLjABAAAAU0QYHCmBaS2uUvbuc5XfY-XTJJCybRun-GHf2kpR6o0?from_tab_name=main" \
  --get_comment yes \
  --get_sub_comment no \
  --max_comments_count_singlenotes 500 \
  --crawler_max_notes_count 50 \
  --max_concurrency_num 1 \
  --save_data_option jsonl \
  --save_data_path "<wiki>/raw/studio/funnel/published/douyin/runs/YYYY-MM-DD"

Use --crawler_max_notes_count 130 or higher for a full-account refresh. If login is needed, let the user scan the QR code in Chrome. Keep each run in a date-stamped directory; never overwrite old runs.

Incremental Merge

After MediaCrawler finishes, run:

python3 <skill_dir>/scripts/update_douyin_archive.py \
  --archive-dir "<wiki>/raw/studio/funnel/published/douyin" \
  --run-dir "<wiki>/raw/studio/funnel/published/douyin/runs/YYYY-MM-DD" \
  --date "YYYY-MM-DD"

The script will:

  • read douyin/jsonl/creator_contents_*.jsonl
  • read douyin/jsonl/creator_comments_*.jsonl when present
  • update manifest/kangjian-douyin-manifest.json
  • create YYYY-MM-DD-kangjian-douyin-douyin-full-ledger.md
  • create YYYY-MM-DD-kangjian-douyin-douyin-review-analysis.md
  • create or refresh latest-kangjian-douyin-summary.md

Use --full-refresh when the run intentionally covers the whole account.

Output Contract

Raw and generated files stay under:

raw/studio/funnel/published/douyin/
  manifest/
    kangjian-douyin-manifest.json
  runs/
    YYYY-MM-DD/
      douyin/jsonl/creator_contents_YYYY-MM-DD.jsonl
      douyin/jsonl/creator_comments_YYYY-MM-DD.jsonl
  YYYY-MM-DD-kangjian-douyin-douyin-full-ledger.md
  YYYY-MM-DD-kangjian-douyin-douyin-review-analysis.md
  latest-kangjian-douyin-summary.md

The manifest is the incrementally maintained index. Each run remains an immutable-ish snapshot for auditability.

Validation

Before reporting completion:

  1. Run wc -l on contents and comments JSONL.
  2. Run the archive update script and report its works, comments, and new_works counts.
  3. Run git diff --check -- raw/studio/funnel/published/douyin.
  4. Mention whether the run was incremental or full refresh, and whether the current visible work count differs from the user's expected count.

Failure Handling

  • If MediaCrawler fails from missing dependencies, run uv sync in the checkout and retry.
  • If browser or login fails, ask the user to approve Chrome/QR login again.
  • If comments are incomplete due to rate limiting, keep the run but say it is partial and rerun later with lower concurrency.
  • If the archive path is outside the current writable workspace, request filesystem approval instead of copying data silently.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.