Spot Trending GitHub Projects and Auto-Log Them to Obsidian

Where this fits

Every Monday morning, you open GitHub Trending, scroll past 50 repos, and mentally filter which ones matter for your current research or side project. Half an hour later you've forgotten half of them. This skill automates that ritual: it fetches the weekly (or daily/monthly) trending list, scores each repo against your saved keywords and domain boosts, and writes a clean Obsidian note into your vault — all with a single trigger phrase and no LLM token spend.

Why agents benefit

• Zero-token execution: The heavy lifting is pure Python (fetch_trending.py), so the agent can run the entire pipeline without calling a language model — ideal for scheduled or high-frequency runs.
• Relevance scoring out of the box: It reuses your existing daily_papers.keywords, negative_keywords, and domain_boost_keywords config, so scoring aligns with your research taste without extra setup.
• Native Obsidian integration: Output lands as a formatted markdown note in your configured GitHubTrending folder — ready for linking, tagging, or further processing.
• Flexible time windows: Parse "周榜", "日榜", or "月榜" from natural language and pass --since weekly|daily|monthly automatically; optional language filter (e.g., "只看 Python") is supported.
• Config-aware and personalizable: It reads shared config plus optional local overrides, so team defaults and personal tweaks stay separate and predictable.

Practical scenarios

Morning research triage

Say "GitHub 周榜" and the agent fetches the weekly trending list, scores each repo against your ML research keywords, and writes a note titled GitHubTrending-2025-04-14.md with a table of repos, stars, and relevance badges. You scan 30 seconds instead of 30 minutes.

Focused language tracking

If you're doing a Python-only deep dive, add "只看 Python" to your trigger. The skill runs --language python and filters out everything else — perfect for keeping your feed narrow before a sprint.

Monthly tech watch

Request "GitHub 月榜" to get a broader view of what gained traction over 30 days. The same scoring pipeline applies, so you can compare monthly momentum against your weekly snapshot and spot emerging trends early.

Add it to your agent workflow

Install the skill globally with:

npx skills add huangkiki/dailypaper-skills --skill github-trending

Then simply say "开始抓 GitHub 榜单 🐙" followed by the period (e.g., "这周"). The agent will confirm the date, run the fetch script, and drop the note into your vault. For a deeper look at how this skill fits into the broader agent-skill ecosystem, check the skill page on OpenAgentSkill.

Compare before adopting

• Quality signals: Check if the repo's fetch_trending.py still works (GitHub's HTML changes often) and whether the scoring logic matches your keyword taxonomy.
• Maintenance freshness: The skill's last commit date and issue tracker tell you if it's actively maintained — stale skills break silently.
• Alternatives: Compare against manual bookmarking, RSS feeds, or other trending scrapers (e.g., gh-trending CLI) for speed and automation depth.
• Workflow fit: Verify the Obsidian folder path and config merge behavior align with your existing vault structure before making it a daily habit.

Why it is worth tracking

With 1,201 stars and an active author (huangkiki), this skill shows real community traction. The zero-token design makes it cheap to run repeatedly, and the config-driven scoring means it improves as you refine your keywords. Worth evaluating now — and revisiting if the author adds new periods or output formats in future releases.