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.
In this repository39 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- logo-generatorSkill
- source-command-kj-helpSkill
- ai-coding-weekly-reportSkill
- ai-conversation-to-wikiSkill
- ai-daily-from-xSkill
- ai-daily-websearchSkill
- ai-learning-loopSkill
- ai-programming-topic-plannerSkill
- ai-reading-coachSkill
- coding-session-to-tutorialSkill
- content-creatorSkill
- content-growth-reviewSkill
- content-illustratorSkill
- course-campaignSkill
- course-generatorSkill
- course-scraperSkill
- english-reading-studioSkill
- extract-conversation-insightsSkill
- git-pushSkill
- humanizer-zhSkill
- industry-best-practicesSkill
- kangjian-skillSkill
- knowledge-mapSkill
- markdown-formatSkill
- paid-content-reviewSkill
- phenomenon-insightSkill
- research-backed-content-loopSkill
- skills-repo-health-checkSkill
- story-pptSkill
- transcript-critical-analysisSkill
- tutorial-guideSkill
- web-slideSkill
- wechat-companionSkill
- wechat-publish-kitSkill
- weekly-digestSkill
- wiki-doc-sinkSkill
- wiki-intelligenceSkill
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
- Distinctive: line 8
Use this skill to update the user's own Douyin archive for:
- Very specific: line 13
Default archive root in the wiki: raw/studio/funnel/published/douyin/
- Very specific: line 17
Use NanmiCoder/MediaCrawler to produce raw Douyin JSONL for works and comments.
- Very specific: line 18
Run scripts/update_douyin_archive.py to merge the run incrementally and generate Markdown reports.
- Very specific: line 20
Do not download media files by default.
- Very specific: line 32
Use an existing MediaCrawler checkout if available.
- Very specific: line 51
Use --crawler_max_notes_count 130 or higher for a full-account refresh.
- Distinctive: line 55
After MediaCrawler finishes, run:
- Very specific: line 67
read douyin/jsonl/creator_comments_*.jsonl when present
- Very specific: line 71
create or refresh latest-kangjian-douyin-summary.md
- Very specific: line 73
Use --full-refresh when the run intentionally covers the whole account.
- Very specific: line 98
Run wc -l on contents and comments JSONL.
- Very specific: line 99
Run the archive update script and report its works, comments, and new_works counts.
- Very specific: line 105
If MediaCrawler fails from missing dependencies, run uv sync in the checkout and retry.
- Distinctive: line 106
If browser or login fails, ask the user to approve Chrome/QR login again.
- Distinctive: line 107
If comments are incomplete due to rate limiting, keep the run but say it is partial and rerun later with lower concurrency.
- Distinctive: line 108
If the archive path is outside the current writable workspace, request filesystem approval instead of copying data silently.
≈ “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.
Files
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:
- Use
NanmiCoder/MediaCrawlerto produce raw Douyin JSONL for works and comments. - Run
scripts/update_douyin_archive.pyto 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_*.jsonlwhen 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.mdThe manifest is the incrementally maintained index. Each run remains an immutable-ish snapshot for auditability.
Validation
Before reporting completion:
- Run
wc -lon contents and comments JSONL. - Run the archive update script and report its
works,comments, andnew_workscounts. - Run
git diff --check -- raw/studio/funnel/published/douyin. - 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 syncin 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.
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
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.

