抓取 AI 大厂官方动态(Anthropic + OpenAI,可继续扩展)→ 中文策展 → 渲染小红书 / 公众号卡片 → GitHub Pages 可视化展示页 → Codex Automation 每天自动更新。 整条流程封装为可安装的 Agent Skill:skills/ai-daily-digest。
FAQ
ai-frontier-daily is a Claude Code plugin with 6 hand-picked skills for content work, indexed on Flowy. Install it with the command on its page. It includes ai-daily-digest, deepdive-xiaohongshu, firsthand-deepdive. Its skills do not fire on their own yet. Request auto-invocation to have Flowy route them as you prompt. Free and open source.
$ npx -y skills add yinjialu/ai-frontier-daily --agent claude-code
Repo: yinjialu/ai-frontier-daily
抓取 AI 大厂官方动态(Anthropic + OpenAI,可继续扩展)→ 中文策展 → 渲染小红书 / 公众号卡片 → GitHub Pages 可视化展示页 → Codex Automation 每天自动更新。
整条流程封装为可安装的 Agent Skill:skills/ai-daily-digest。
https://<你的用户名>.github.io/ai-frontier-daily/
手机打开即可浏览当天卡片:顶部 Anthropic / OpenAI 切换,各厂商独立主题;小红书 6 张竖图直接发笔记,公众号封面 + 正文长图手动建草稿。
# 装到当前项目
npx skills add yinjialu/ai-frontier-daily
# 或装到全局(所有项目可用)
npx skills add yinjialu/ai-frontier-daily --skill ai-daily-digest -g -a claude-code -y
安装后在 Claude Code 里说「做今天的 Anthropic 动态卡片」「更新每日 AI 速递」即可触发。
Skill 用法详见 skills/ai-daily-digest/SKILL.md。
node render.js 渲染 → 提交触发 Pages → PushNotification 推手机。skills/ai-daily-digest/SKILL.md,遍历 run.py --vendors 返回的全部厂商,抓取一手动态、
中文策展、Playwright 渲染,提交 data/ 与 output/ 到 Codex/daily-YYYY-MM-DD。.github/workflows/daily.yml):
只做信源健康探测和当天产出校验,异常时开 Issue,不负责生成内容。index.html 读
output/index.json 按厂商展示当天卡片。gh repo create ai-frontier-daily --public --source=. --remote=origin --push
# 开启 Pages:从 main 分支根目录部署
gh api -X POST repos/$(gh api user -q .login)/ai-frontier-daily/pages \
-f "source[branch]=main" -f "source[path]=/" || \
echo "→ 改用网页:Settings → Pages → Deploy from a branch → main / (root)"
# 配置 Actions 密钥(live 模式调 Claude 用;切勿写进代码)
gh secret set ANTHROPIC_API_KEY --body "sk-ant-你的key"
# 早报生成由 Codex Automation 每天 22:00 自动预制次日刊;Actions daily.yml 只做哨兵校验。
pip install -r skills/ai-daily-digest/requirements.txt
npm i playwright && npx playwright install chromium
python skills/ai-daily-digest/run.py # mock,无需联网/密钥
python -m http.server 8000 # 打开 http://localhost:8000/index.html
data/ 与 output/ 默认写到当前目录(可用环境变量 DIGEST_OUT 指定)。
ANTHROPIC_API_KEY 只放 Actions Secrets,绝不进仓库。.env.example
.github/
workflows/
auto-merge-daily.yml
auto-merge-firsthand.yml
claude.yml
daily.yml
deepseek.yml
pages-deploy.yml
.gitignore
.nojekyll
AGENTS.md
CLAUDE.md
conftest.py
content/
deepdive/
2026-06-24_introducing-claude-tag/
assets/
claude-tag-illustrations/
01-per-user-permission-risk.png
02-per-channel-agent-identity.png
cover-wechat-21x9.png
cover-xiaohei-agent-identity.png
draft.md
xhs.json
2026-06-29_gpt-5-6-model-family/
assets/
cover-prompt.md
cover-wechat-21x9.png
cover-xiaohei-gpt56-model-family-alt.png
cover-xiaohei-gpt56-model-family.png
draft.md
2026-07-01_getting-started-with-loops/
assets/
cover-wechat-21x9.png
cover-wechat-xiaohei-runtime.png
getting-started-with-loops-illustrations/
01-prompt-to-loop.png
02-stop-condition.png
03-agent-infrastructure.png
draft.md
xhs.json
2026-07-02_redeploying-fable-5/
assets/
xiaohei-illustrations/
01-model-gate.png
02-jailbreak-scale.png
03-release-governance.png
draft.md
illustration-plan.md
wechat-meta.md
xhs.json
2026-07-08_claude-cowork-knowledge-work/
assets/
cover-knowledge-work.png
wechat-1x1-cover.png
wechat-21x9-cover.png
draft.md
wechat-meta.md
xhs-caption.txt
xhs-images/
00-cover.jpg
01-card.jpg
02-card.jpg
03-card.jpg
04-card.jpg
05-card.jpg
06-card.jpg
07-card.jpg
xhs.json
2026-07-08_making-of-claude-code/
assets/
claude-code-logo-slate.png
claude-icon-rounded.png
claude-spark-clay.png
cover-wechat-900x383.jpg
making-of-claude-code-og.gif
draft.md
image-options.md
xhs-images/
xhs-01-cover.png
xhs-02-core-claim.png
xhs-03-coding-feedback.png
xhs-04-prototype.png
xhs-05-cli.png
xhs-06-product-loop.png
xhs-07-takeaway.png
xhs.json
2026-07-09_claude-code-model-effort/
assets/
cover-model-effort.png
wechat-1x1-cover.png
wechat-21x9-cover.png
draft.md
wechat-meta.md
xhs-caption.txt
xhs-images/
00-cover.jpg
01-card.jpg
02-card.jpg
03-card.jpg
04-card.jpg
05-card.jpg
06-card.jpg
07-card.jpg
xhs.json
2026-07-11_cognition-fable-5/
assets/
cover-cognition-fable-5-xhs.png
cover-cognition-fable-5.png
draft.md
image-plan.md
wechat-meta.md
xhs.json
data/
anthropic/
2026-05-31.json
2026-06-01.json
2026-06-02.json
2026-06-03.json
2026-06-04.json
2026-06-05.json
2026-06-06.json
2026-06-07.json
2026-06-08.json
2026-06-09.json
2026-06-10.json
2026-06-11.json
2026-06-12.json
2026-06-13.json
2026-06-14.json
2026-06-15.json
2026-06-16.json
2026-06-17.json
2026-06-18.json
2026-06-19.json
2026-06-20.json
2026-06-21.json
2026-06-22.json
2026-06-23.json
2026-06-24.json
2026-06-25.json
2026-06-26.json
2026-06-27.json
2026-06-29.json
2026-06-30.json
2026-07-01.json
2026-07-02.json
2026-07-03.json
2026-07-04.json
2026-07-05.json
2026-07-07.json
2026-07-08.json
2026-07-09.json
2026-07-11.json
2026-07-14.json
2026-07-15.json
2026-07-17.json
2026-07-18.json
2026-07-19.json
2026-07-21.json
2026-07-22.json
cn/
2026-06-16.json
2026-06-17.json
2026-06-18.json
2026-06-19.json
2026-06-21.json
2026-06-22.json
2026-06-23.json
2026-06-24.json
2026-06-25.json
2026-06-26.json
2026-06-27.json
2026-06-28.json
2026-06-29.json
2026-06-30.json
2026-07-01.json
2026-07-02.json
2026-07-03.json
2026-07-05.json
2026-07-06.json
2026-07-07.json
2026-07-08.json
2026-07-09.json
2026-07-17.json
2026-07-19.json
2026-07-21.json
2026-07-24.json
firsthand/
firsthand-state.json
2026-07-10/
.heartbeat
2026-07-12/
.heartbeat
2026-07-13/
.heartbeat
2026-07-14/
.heartbeat
2026-07-15/
.heartbeat
2026-07-19/
.heartbeat
anthropic-engineering/
01-contextual-retrieval.md
02-building-effective-agents.md
03-swe-bench-sonnet.md
04-claude-think-tool.md
05-claude-code-best-practices.md
06-multi-agent-research-system.md
07-desktop-extensions.md
08-writing-tools-for-agents.md
09-a-postmortem-of-three-recent-issues.md
10-effective-context-engineering-for-ai-agents.md
11-equipping-agents-for-the-real-world-with-agent-skills.md
12-claude-code-sandboxing.md
13-code-execution-with-mcp.md
14-advanced-tool-use.md
15-effective-harnesses-for-long-running-agents.md
16-demystifying-evals-for-ai-agents.md
17-AI-resistant-technical-evaluations.md
18-building-c-compiler.md
19-infrastructure-noise.md
20-eval-awareness-browsecomp.md
21-harness-design-long-running-apps.md
22-claude-code-auto-mode.md
23-managed-agents.md
24-april-23-postmortem.md
25-how-we-contain-claude.md
anthropic-news/
01-chris-olah-pope-leo-encyclical.md
02-claude-opus-4-8.md
03-expanding-project-glasswing.md
04-services-track-partner-hub.md
05-AI-enabled-cyber-threats-mitre-attack.md
06-claude-fable-5-mythos-5.md
07-claude-corps.md
08-dxc-anthropic-alliance.md
09-fable-mythos-access.md
10-anthropic-public-record.md
11-tcs-anthropic-partnership.md
12-seoul-office-partnerships-korean-ai-ecosystem.md
13-introducing-claude-tag.md
14-claude-science-ai-workbench.md
14-claude-sonnet-5.md
14-redeploying-fable-5.md
15-redeploying-fable-5.md
16-fable-safeguards-jailbreak-framework.md
17-reflect-with-claude.md
18-hard-questions.md
19-ust-claude.md
20-ben-bernanke.md
21-claude-for-teachers.md
22-canadian-ai-research.md
23-rare-disease-research-grants.md
24-donation-public-first-action.md
25-economic-futures-research-fund-agenda.md
26-anthropic-economic-index-connector.md
27-claude-opus-5.md
28-position-open-weights-models.md
29-cognizant-anthropic.md
30-investigating-incidents-cybersecurity-evals.md
anthropic-research/
01-natural-language-autoencoders.md
02-teaching-claude-why.md
03-glasswing-initial-update.md
04-exploit-evals.md
05-coding-agents-social-sciences.md
06-attack-navigator.md
07-making-claude-a-chemist.md
08-agents-in-biology.md
09-n-days.md
10-claude-code-expertise.md
11-project-fetch-phase-two.md
12-economic-index-june-2026-report.md
13-off-switch-dual-use.md
14-claude-plays-robotics.md
15-claude-values-models-languages.md
16-how-canada-uses-claude.md
17-project-pilot.md
18-discovering-cryptographic-weaknesses.md
claude-blog/
01-harnessing-claudes-intelligence.md
02-claude-managed-agents.md
03-preparing-your-security-program-for-ai-accelerated-offense.md
04-claude-code-desktop-redesign.md
05-claude-managed-agents-memory.md
06-connectors-for-everyday-life.md
07-claude-managed-agents-updates.md
08-new-in-claude-managed-agents.md
09-how-anthropic-enables-self-service-data-analytics-with-claude.md
10-how-anthropic-uses-claude-gtm-engineering.md
11-the-claude-cowork-product-guide.md
12-claude-for-foundation-models.md
13-observability-for-developers-building-connectors.md
14-whats-new-in-claude-managed-agents.md
15-building-with-claude-managed-agents.md
16-meet-the-winners-of-built-with-opus-4-7-claude-code-hackathon.md
17-claude-design-stays-on-brand-for-daily-work.md
18-meet-the-winners-of-our-claude-opus-4-8-build-day-hackathon.md
19-workload-identity-federation.md
20-artifacts-in-claude-code.md
21-enterprise-managed-auth.md
22-steering-claude-code-skills-hooks-rules-subagents-and-more.md
23-the-full-claude-desktop-experience-on-aws-google-cloud-and-microsoft-foundry.md
24-agent-identity-access-model.md
24-building-effective-human-agent-teams.md
25-getting-started-with-loops.md
25-introducing-the-claude-apps-gateway.md
26-claude-in-microsoft-foundry.md
27-giving-admins-more-visibility-and-control-over-claude-usage-and-spend.md
28-bringing-claude-code-and-claude-cowork-to-government.md
29-claude-model-and-effort-level-in-claude-code.md
30-cowork-web-mobile.md
31-how-people-are-using-claude-cowork.md
32-working-at-the-frontier-how-thomson-reuters-builds-ai-for-high--stakes-professional-work.md
33-how-anthropics-marketing-operations-team-uses-claude-cowork-to-automate-reporting-and-campaign-builds.md
34-working-at-the-frontier-how-cognition-trusts-claude-fable-5-to-work-through-the-night.md
35-working-at-the-frontier-how-hebbia-builds-ai-for-financial-diligence-that-cant-miss-a-detail.md
36-working-at-the-frontier-why-base44-trusts-claude-fable-5-with-their-most-challenging-engineering-work.md
37-working-with-claude-fable-5-in-claude-cowork.md
38-ai-code-migration.md
38-working-at-the-frontier-cursor.md
39-ciso-guide-to-agentic-ai.md
40-how-anthropic-secures-its-ai-native-software-development-lifecycle.md
40-working-at-the-frontier-rakuten.md
41-how-datadog-built-a-universal-machine-tool-for-claude-code.md
42-building-verification-loops-in-claude-code-with-skills.md
43-how-outtake-built-a-cyber-investigator-on-claude.md
44-think-through-hard-problems-in-voice-mode.md
45-four-role-based-claude-certifications.md
46-how-the-product-designer-who-built-claude-design-uses-it-to-explore-ideas-before-building-them.md
47-the-new-rules-of-context-engineering-for-claude-5-generation-models.md
48-claude-models-explained-choosing-the-best-model-for-your-use-case.md
49-bringing-mcp-2026-07-28-to-claude.md
deepmind-blog/
01-introducing-computer-use-in-gemini-3-5-flash.md
02-start-building-with-nano-banana-2-lite-and-gemini-omni-flash.md
03-google-deepmind-and-a24-announce-first-of-its-kind-research-partnership.md
04-empowering-indias-next-generation-of-innovators-with-atl-saathi.md
05-our-approach-to-bioresilience.md
06-introducing-gemini-3-5-flash-cyber.md
07-introducing-gemini-36-flash-35-flash-lite-and-35-flash-cyber.md
08-accelerating-the-frontiers-of-scientific-discovery-googles-40m-commitment-to-the-genesis-mission.md
09-introducing-gemini-3-6-flash-3-5-flash-lite-and-3-5-flash-cyber.md
10-gemini-robotics-2-brings-whole-body-intelligence-to-robots.md
11-were-launching-lyria-35-in-google-flow-music-with-advances-across-musicality-lyrics-vocals-and-creative-control.md
12-gemini-robotics-er-2-powering-robotics-with-video-understanding-task-orchestration-and-multi-robot-collaboration.md
feed.xml
google-gemini/
01-find-job-with-google-ai-tools.md
01-gemini-study-notebooks.md
01-gemini-tips-for-parents.md
01-introducing-computer-use-gemini-3-5-flash.md
02-gemini-help-avoid-jetlag.md
03-gemini-omni-flash-nano-banana-2-lite.md
03-gemini-spark-updates-june-2026.md
03-personal-intelligence-nano-banana-us-expansion.md
04-take-notes-for-me.md
05-google-ai-updates-june-2026.md
06-coffee-shop-gemini-features.md
07-how-to-make-gemini-study-notebooks.md
08-gemini-southeast-asia-report-2026.md
09-launch-business-with-gemini.md
09-summer-productivity-tips-google-ai-tools.md
10-gemini-3-6-flash-3-5-flash-lite-3-5-flash-cyber.md
11-gemini-live-camera-how-to.md
12-galaxy-unpacked-2026.md
13-using-gemini-to-manage-farm.md
14-speak-naturally-gemini-app-mac-os.md
15-gemini-spark-updates-july-2026.md
16-gemini-robotics-er-2.md
17-glanceboard-gemini-flash-nano-banana.md
18-gemini-drop-july-2026.md
google-models-research/
01-introducing-computer-use-gemini-3-5-flash.md
02-gemini-omni-flash-nano-banana-2-lite.md
02-google-ai-updates-june-2026.md
03-technology-global-crisis-resilience.md
04-firesat-satellites.md
05-alphaevolve-on-cloud.md
06-reconstructing-peles-lost-goal.md
07-gemini-3-6-flash-3-5-flash-lite-3-5-flash-cyber.md
08-understanding-the-ai-economy.md
09-using-gemini-to-manage-farm.md
10-lyria-3-5.md
11-gemini-robotics-er-2.md
12-glanceboard-gemini-flash-nano-banana.md
index.json
openai-news/
01-openai-broadcom-jalapeno-inference-chip.md
02-how-agents-are-transforming-work.md
03-previewing-gpt-5-6-sol.md
04-how-chatgpt-adoption-has-expanded.md
04-introducing-genebench-pro.md
05-case-studies.md
05-core-dump-epidemiology-data-infrastructure-bug.md
06-australian-payments-plus.md
07-government-national-security-partnerships.md
08-separating-signal-from-noise-coding-evaluations.md
... 1600 moreShowing a partial view of a very large repo.
© 2026 Flowy · Free and open source
Built for Claude Code · Not affiliated with Anthropic