Daily Briefing Skill
Aggregates 5 news sources into a formatted daily digest with summaries. Designed for Telegram push via OpenClaw cron.
Sources
| # | Source | API / Method |
|---|---|---|
| 1 | Readhub 早报 | https://api.readhub.cn/daily — Chinese tech news with full summaries |
| 2 | Karpathy RSS | 92 HN popular blogs OPML, concurrent fetch, latest posts with descriptions |
| 3 | GitHub Trending | Scrape github.com/trending + GitHub API for repo descriptions & stars |
| 4 | Hacker News | Firebase API hacker-news.firebaseio.com/v0/topstories.json |
| 5 | V2EX 热帖 | https://www.v2ex.com/api/topics/hot.json |
Quick Start
Run once (on-demand)
python3 scripts/daily-briefing.py 2>/dev/nullOutput is Markdown-formatted, suitable for direct Telegram send.
Set up daily cron (08:00 CST)
openclaw cron add \
--name "每日资讯早报" \
--cron "0 8 * * *" \
--session isolated \
--announce \
--channel telegram \
--thinking low \
--message "运行 python3 <workspace>/scripts/daily-briefing.py 2>/dev/null 并将完整输出分段发送到 Telegram 主群"Replace <workspace> with actual path (e.g. /Users/you/.openclaw/workspace-main).
Send output to Telegram (split by section)
The output is ~200 lines. Split into 4 messages by section (📰 / 🔖 / ⭐ / 🔥+💬) to stay under Telegram's 4096-char limit.
Script Details
File: scripts/daily-briefing.py
- Pure Python stdlib only — no pip dependencies
- Concurrent fetching: 25 workers for RSS feeds, 10 for HN items, 8 for GitHub API
- Total runtime: ~25 seconds
- Karpathy OPML source:
https://gist.githubusercontent.com/emschwartz/e6d2bf860ccc367fe37ff953ba6de66b/raw/hn-popular-blogs-2025.opml
Customization
Change number of items per section — edit the n= parameter in each get_*() call in main():
readhub = get_readhub(8) # default 8
rss = get_karpathy_rss(10) # default 10
gh = get_github_trending(8) # default 8
hn = get_hn_top(8) # default 8
v2 = get_v2ex_hot(8) # default 8Disable a section — comment out its block in main().
Change RSS feed list — the script dynamically fetches the OPML from GitHub Gist on each run. To use a static local list instead, replace load_opml_feeds() return value with a hardcoded list.