Qiaomu Ai Radar

# Runtime state
data/*
!data/.gitkeep
logs/

# Python caches
__pycache__/
*.py[cod]

# Local/editor noise
.DS_Store
.vscode/
.idea/

# Agent runtime counters
.call_count
.circuit_breaker_history
.circuit_breaker_state
.exit_signals
.last_reset
progress.json
status.json

# Databases and logs
*.db
*.db-shm
*.db-wal
*.log

<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE plist PUBLIC "-//Apple//DTD PLIST 1.0//EN" "http://www.apple.com/DTDs/PropertyList-1.0.dtd">
<plist version="1.0">
<dict>
    <key>Label</key>
    <string>com.user.qiaomu-ai-radar</string>

    <key>ProgramArguments</key>
    <array>
        <string>/usr/bin/python3</string>
        <string>/Users/YOUR_USERNAME/.agents/skills/qiaomu-ai-radar/scripts/run_radar.py</string>
        <string>--mode</string>
        <string>fetch</string>
    </array>

    <key>EnvironmentVariables</key>
    <dict>
        <key>PATH</key>
        <string>/Users/YOUR_USERNAME/.local/node/bin:/opt/homebrew/bin:/usr/local/bin:/usr/bin:/bin</string>
    </dict>

    <key>StartCalendarInterval</key>
    <dict>
        <key>Hour</key>
        <integer>9</integer>
        <key>Minute</key>
        <integer>0</integer>
    </dict>

    <key>StandardOutPath</key>
    <string>/Users/YOUR_USERNAME/.agents/skills/qiaomu-ai-radar/data/cron.log</string>

    <key>StandardErrorPath</key>
    <string>/Users/YOUR_USERNAME/.agents/skills/qiaomu-ai-radar/data/cron.error.log</string>
</dict>
</plist>

{
  "blacklist": {
    "keywords": {
      "title": [
        "lua",
        "lisp",
        "haskell",
        "rust compiler",
        "UEFI",
        "GPIB",
        "FPGA",
        "HBM",
        "scheduler",
        "garbage collection",
        "TCP_NODELAY",
        "HTTP caching",
        "PostgreSQL",
        "NixOS",
        "LoongArch",
        "LSP support",
        "fined",
        "law blocked",
        "spying",
        "privacy",
        "router",
        "font",
        "typeface",
        "runway design",
        "billiards",
        "μs off",
        "redactions",
        "charts in slides",
        "lockfiles",
        "SBOM",
        "terminal emulator",
        "ghostty",
        "wezterm",
        "alacritty",
        "AI safety",
        "AI governance",
        "AI regulation",
        "situational awareness",
        "effective altruism",
        "executive briefing",
        "corporate AI adoption",
        "VC funding round",
        "series A funding",
        "series B funding",
        "fundraising round"
      ],
      "url": [
        "twitter.com",
        "x.com"
      ],
      "summary": [
        "sponsor"
      ]
    },
    "sources": {
      "Wait But Why": {
        "exclude_url_patterns": [
          "#comments$"
        ]
      }
    }
  },
  "whitelist": {
    "keywords": {
      "title": [
        "AI",
        "LLM",
        "GPT",
        "Claude",
        "transformer",
        "machine learning",
        "engineering blog",
        "architecture",
        "design philosophy",
        "bootstrapped",
        "startup",
        "senior engineer",
        "coding",
        "software design",
        "MCP",
        "tool use",
        "WebMCP",
        "AGENTS.md",
        "from scratch",
        "interactive",
        "explainer",
        "SVG",
        "visualization"
      ]
    }
  },
  "rules": {
    "apply_order": "blacklist_first",
    "case_sensitive": false,
    "comment": "黑名单优先：先过滤黑名单，再检查白名单"
  }
}

{
  "version": "1.0",
  "description": "qiaomu-ai-radar 输出路径配置 - 所有用户可见文件统一保存到知识库",

  "_comment_output_base": "基础输出目录 - 所有生成的文件都保存到此处",
  "output_base_dir": "~/乔木新知识库/20-29 学习/21 素材索引",

  "_comment_files": "日常生成的文件名",
  "files": {
    "content_log": "content_log.md",
    "recommendations": "recommendations.md"
  },

  "_comment_index": "索引文件的命名模板",
  "index_files": {
    "weekly_template": "{year}-第{week}周.md",
    "starred": "⭐️收藏精选.md"
  },

  "_comment_internal": "内部状态文件仍保留在技能目录",
  "internal_files": {
    "state": "data/state.json",
    "database": "data/articles.db"
  },

  "_comment_feishu": "飞书同步配置 - 把本周索引镜像到云空间目录",
  "feishu": {
    "enabled": true,
    "cli": "/opt/homebrew/bin/lark-cli",
    "identity": "user",
    "folder_name": "乔木AI雷达",
    "parent_folder_token": "",
    "state_file": "data/feishu_sync.json"
  }
}

{
  "version": "1.0",
  "description": "渠道偏好配置 - 喜欢的渠道排在前面，带⭐标记",

  "favorite_sources": [
    "NLP Newsletter (Elvis Saravia)"
  ],

  "muted_sources": [],

  "_comment": "favorite_sources: 偏好渠道名（对应 sources.json 中的 name 字段），排在前面且带⭐标记。muted_sources: 静音渠道，采集但折叠显示。"
}

{
  "sources": [
    {
      "id": "james-clear",
      "name": "James Clear 3-2-1",
      "url": "https://jamesclear.com/3-2-1",
      "type": "newsletter_archive",
      "parser": "jamesclear",
      "enabled": true,
      "check_interval_hours": 168,
      "last_checked": "2026-06-02T01:00:19.870362+00:00"
    },
    {
      "id": "bensbites",
      "name": "Ben's Bites",
      "url": "https://www.bensbites.com/archive",
      "type": "newsletter_archive",
      "parser": "bensbites",
      "enabled": true,
      "check_interval_hours": 24,
      "limit": 10,
      "last_checked": "2026-06-02T01:00:19.870237+00:00"
    },
    {
      "id": "tldrai",
      "name": "TLDR AI",
      "url": "https://tldr.tech/api/latest/ai",
      "type": "newsletter",
      "parser": "tldrai",
      "enabled": true,
      "check_interval_hours": 24,
      "limit": 10,
      "last_checked": "2026-06-02T01:00:19.870440+00:00"
    },
    {
      "id": "importai",
      "name": "Import AI",
      "url": "https://importai.substack.com/feed",
      "type": "newsletter_rss",
      "parser": "importai",
      "enabled": true,
      "check_interval_hours": 168,
      "limit": 10,
      "last_checked": "2026-06-02T01:00:19.870123+00:00"
    },
    {
      "id": "nlp-elvis",
      "name": "NLP Newsletter (Elvis Saravia)",
      "url": "https://nlp.elvissaravia.com/feed",
      "type": "newsletter_rss",
      "parser": "importai",
      "enabled": true,
      "check_interval_hours": 168,
      "limit": 5,
      "last_checked": "2026-06-02T01:00:19.870359+00:00"
    },
    {
      "id": "interconnects",
      "name": "Interconnects (Nathan Lambert)",
      "url": "https://www.interconnects.ai/feed",
      "type": "newsletter_rss",
      "parser": "importai",
      "enabled": true,
      "check_interval_hours": 24,
      "limit": 5,
      "last_checked": "2026-06-02T01:00:19.870343+00:00"
    },
    {
      "id": "huggingface",
      "name": "Hugging Face Papers",
      "url": "https://huggingface.co/papers",
      "type": "aggregator",
      "parser": "huggingface",
      "enabled": true,
      "check_interval_hours": 24,
      "limit": 15,
      "min_votes": 30,
      "last_checked": "2026-06-02T01:00:19.870312+00:00"
    },
    {
      "id": "waitbutwhy",
      "name": "Wait But Why",
      "url": "https://waitbutwhy.com",
      "type": "blog",
      "parser": "waitbutwhy",
      "enabled": true,
      "check_interval_hours": 168,
      "last_checked": "2026-06-02T01:00:19.869956+00:00"
    },
    {
      "id": "hackernews",
      "name": "Hacker News",
      "url": "https://news.ycombinator.com/",
      "type": "aggregator",
      "parser": "hackernews",
      "enabled": false,
      "_note": "用户要求禁用，信息噪音太大",
      "check_interval_hours": 24,
      "limit": 8,
      "last_checked": "2026-04-18T08:45:50.442522+00:00"
    },
    {
      "id": "producthunt",
      "name": "Product Hunt",
      "url": "https://www.producthunt.com",
      "type": "aggregator",
      "parser": "producthunt",
      "enabled": true,
      "check_interval_hours": 24,
      "limit": 10,
      "last_checked": "2026-06-02T01:00:19.870368+00:00"
    },
    {
      "id": "yuanchaofa",
      "name": "袁超发技术博客",
      "url": "https://yuanchaofa.com/rss.xml",
      "type": "blog",
      "parser": "yuanchaofa",
      "enabled": true,
      "check_interval_hours": 168,
      "limit": 10,
      "last_checked": "2026-06-02T01:00:19.870481+00:00"
    },
    {
      "id": "brainfood",
      "name": "Farnam Street Brain Food",
      "url": "https://fs.blog/brain-food/",
      "type": "newsletter_rss",
      "parser": "brainfood",
      "enabled": true,
      "check_interval_hours": 168,
      "limit": 10,
      "last_checked": "2026-06-02T01:00:19.870514+00:00"
    },
    {
      "id": "austinkleon",
      "name": "Austin Kleon",
      "url": "https://austinkleon.substack.com/archive",
      "type": "newsletter_rss",
      "parser": "austinkleon",
      "enabled": true,
      "check_interval_hours": 168,
      "limit": 10,
      "last_checked": "2026-06-02T01:00:19.870584+00:00"
    },
    {
      "id": "paulgraham",
      "name": "Paul Graham Essays",
      "url": "https://paulgraham.com/articles.html",
      "type": "blog",
      "parser": "paulgraham",
      "enabled": true,
      "check_interval_hours": 168,
      "limit": 15,
      "last_checked": "2026-06-02T01:00:19.871368+00:00"
    },
    {
      "id": "lexfridman",
      "name": "Lex Fridman Podcast",
      "url": "https://lexfridman.com/feed/podcast/",
      "type": "podcast",
      "parser": "lexfridman",
      "enabled": true,
      "check_interval_hours": 168,
      "limit": 10,
      "last_checked": "2026-06-02T01:00:19.871914+00:00"
    },
    {
      "id": "cognitiverevolution",
      "name": "Cognitive Revolution",
      "url": "https://feeds.megaphone.fm/RINTP3108857801",
      "type": "podcast",
      "parser": "cognitiverevolution",
      "enabled": false,
      "_note": "用户明确表示不感兴趣(AI safety/policy方向)",
      "check_interval_hours": 168,
      "limit": 10,
      "last_checked": "2026-03-02T03:15:28.938078+00:00"
    },
    {
      "id": "hours80k",
      "name": "80,000 Hours Podcast",
      "url": "https://feeds.transistor.fm/80000-hours-podcast",
      "type": "podcast",
      "parser": "hours80k",
      "enabled": false,
      "check_interval_hours": 168,
      "limit": 10,
      "last_checked": "2026-03-02T03:15:31.617653+00:00"
    },
    {
      "id": "latentspace",
      "name": "Latent Space Podcast",
      "url": "https://www.latent.space/feed",
      "type": "podcast",
      "parser": "latentspace",
      "enabled": true,
      "check_interval_hours": 168,
      "limit": 10,
      "last_checked": "2026-06-02T01:00:19.871242+00:00"
    },
    {
      "id": "oneusefulthing",
      "name": "One Useful Thing (Ethan Mollick)",
      "url": "https://www.oneusefulthing.org/feed",
      "type": "newsletter_rss",
      "parser": "oneusefulthing",
      "enabled": true,
      "check_interval_hours": 24,
      "limit": 10,
      "last_checked": "2026-06-02T01:00:19.871296+00:00"
    },
    {
      "id": "hackernoon-lifehacking",
      "name": "HackerNoon Life Hacking",
      "url": "https://hackernoon.com/c/life-hacking",
      "type": "aggregator",
      "parser": "hackernoon",
      "enabled": false,
      "_note": "16篇0阅读",
      "check_interval_hours": 24,
      "limit": 10,
      "last_checked": "2026-03-02T03:15:38.332150+00:00"
    },
    {
      "id": "hackernoon-writing",
      "name": "HackerNoon Writing",
      "url": "https://hackernoon.com/c/writing",
      "type": "aggregator",
      "parser": "hackernoon",
      "enabled": false,
      "_note": "18篇0阅读",
      "check_interval_hours": 24,
      "limit": 10,
      "last_checked": "2026-03-02T03:15:40.544115+00:00"
    },
    {
      "id": "hackernoon-pm",
      "name": "HackerNoon Product Management",
      "url": "https://hackernoon.com/c/product-management",
      "type": "aggregator",
      "parser": "hackernoon",
      "enabled": false,
      "_note": "15篇0阅读",
      "check_interval_hours": 24,
      "limit": 10,
      "last_checked": "2026-03-02T03:15:42.784834+00:00"
    },
    {
      "id": "kdnuggets-ai",
      "name": "KDnuggets AI",
      "url": "https://www.kdnuggets.com/tag/artificial-intelligence",
      "type": "aggregator",
      "parser": "kdnuggets",
      "enabled": false,
      "_note": "92篇0阅读，纯噪音",
      "check_interval_hours": 24,
      "limit": 10,
      "last_checked": "2026-03-02T03:15:46.141494+00:00"
    },
    {
      "id": "scotthyoung",
      "name": "Scott H Young",
      "url": "https://www.scotthyoung.com/blog/articles/",
      "type": "blog",
      "parser": "scotthyoung",
      "enabled": true,
      "check_interval_hours": 168,
      "limit": 10,
      "last_checked": "2026-06-02T01:00:19.871432+00:00"
    },
    {
      "id": "readwise-wise",
      "name": "Readwise Wise",
      "url": "https://wise.readwise.io/",
      "type": "newsletter",
      "parser": "readwise",
      "enabled": true,
      "check_interval_hours": 168,
      "limit": 10,
      "last_checked": "2026-06-02T01:00:19.872010+00:00"
    },
    {
      "id": "dcthemedian",
      "name": "DC The Median",
      "url": "https://dcthemedian.substack.com/",
      "type": "newsletter_rss",
      "parser": "dcthemedian",
      "enabled": false,
      "_note": "10篇0阅读",
      "check_interval_hours": 24,
      "limit": 3,
      "last_checked": "2026-03-02T03:15:55.657762+00:00"
    },
    {
      "id": "markmcneilly",
      "name": "Mark McNeilly",
      "url": "https://markmcneilly.substack.com/",
      "type": "newsletter_rss",
      "parser": "markmcneilly",
      "enabled": false,
      "_note": "12篇0阅读",
      "check_interval_hours": 24,
      "limit": 3,
      "last_checked": "2026-03-02T03:15:58.226235+00:00"
    },
    {
      "id": "businessanalytics",
      "name": "Business Analytics",
      "url": "https://businessanalytics.substack.com/",
      "type": "newsletter_rss",
      "parser": "businessanalytics",
      "enabled": false,
      "check_interval_hours": 24,
      "limit": 3,
      "last_checked": "2026-01-11T12:15:01.346047+00:00"
    },
    {
      "id": "whytryai",
      "name": "Why Try AI",
      "url": "https://www.whytryai.com/",
      "type": "newsletter_rss",
      "parser": "whytryai",
      "enabled": true,
      "check_interval_hours": 24,
      "limit": 3,
      "last_checked": "2026-06-02T01:00:19.871409+00:00"
    },
    {
      "id": "therundown",
      "name": "The Rundown AI",
      "url": "https://rss.beehiiv.com/feeds/2R3C6Bt5wj.xml",
      "type": "newsletter_rss",
      "parser": "therundown",
      "enabled": false,
      "_note": "39篇0阅读，与TLDR AI重复覆盖",
      "check_interval_hours": 24,
      "limit": 3,
      "last_checked": "2026-03-02T03:16:02.943836+00:00"
    },
    {
      "id": "theneuron",
      "name": "The Neuron Daily",
      "url": "https://rss.beehiiv.com/feeds/N4eCstxvgX.xml",
      "type": "newsletter_rss",
      "parser": "theneuron",
      "enabled": false,
      "_note": "59篇0阅读，与TLDR AI/Rundown AI重复",
      "check_interval_hours": 24,
      "limit": 3,
      "last_checked": "2026-03-02T03:16:05.367977+00:00"
    },
    {
      "id": "aileadershipedge",
      "name": "AI Leadership Edge",
      "url": "https://theaileadershipedge.substack.com/",
      "type": "newsletter_rss",
      "parser": "aileadershipedge",
      "enabled": false,
      "check_interval_hours": 24,
      "limit": 3,
      "last_checked": "2026-03-02T03:16:07.549035+00:00"
    },
    {
      "id": "chinai",
      "name": "ChinAI Newsletter",
      "url": "https://chinai.substack.com/",
      "type": "newsletter_rss",
      "parser": "chinai",
      "enabled": true,
      "check_interval_hours": 168,
      "limit": 3,
      "last_checked": "2026-06-02T01:00:19.871336+00:00"
    },
    {
      "id": "memia",
      "name": "Memia (Ben Reid)",
      "url": "https://memia.substack.com/",
      "type": "newsletter",
      "parser": "memia",
      "enabled": false,
      "_note": "15篇0阅读",
      "check_interval_hours": 168,
      "limit": 5,
      "last_checked": "2026-03-02T03:16:12.172777+00:00"
    },
    {
      "id": "ai2roi",
      "name": "AI to ROI",
      "url": "https://ai2roi.substack.com/",
      "type": "newsletter",
      "parser": "ai2roi",
      "enabled": false,
      "check_interval_hours": 24,
      "limit": 5,
      "last_checked": "2026-03-02T03:16:14.612613+00:00"
    },
    {
      "id": "natesnewsletter",
      "name": "Nate's Newsletter",
      "url": "https://natesnewsletter.substack.com/",
      "type": "newsletter",
      "parser": "natesnewsletter",
      "enabled": false,
      "_note": "企业AI战略方向，用户明确不感兴趣",
      "check_interval_hours": 24,
      "limit": 5,
      "last_checked": "2026-03-02T03:16:16.438555+00:00"
    },
    {
      "id": "aichangeseverything",
      "name": "AI Changes Everything",
      "url": "https://patmcguinness.substack.com/",
      "type": "newsletter",
      "parser": "aichangeseverything",
      "enabled": false,
      "_note": "12篇0阅读",
      "check_interval_hours": 24,
      "limit": 5,
      "last_checked": "2026-03-02T03:16:18.571802+00:00"
    },
    {
      "id": "dankoe",
      "name": "Dan Koe Letters",
      "url": "https://thedankoe.com/letters/",
      "type": "newsletter_archive",
      "parser": "dankoe",
      "enabled": true,
      "check_interval_hours": 168,
      "limit": 10,
      "last_checked": "2026-06-02T01:00:19.871474+00:00"
    },
    {
      "id": "shewritesai",
      "name": "She Writes AI",
      "url": "https://shewritesai.substack.com/feed",
      "type": "newsletter_rss",
      "parser": "importai",
      "enabled": false,
      "_note": "11篇0阅读",
      "check_interval_hours": 24,
      "limit": 5,
      "description": "AI writing and content creation insights",
      "last_checked": "2026-03-02T03:16:23.724089+00:00"
    },
    {
      "id": "lolitataub",
      "name": "Lolita Taub",
      "url": "https://lolitataub.substack.com/feed",
      "type": "newsletter_rss",
      "parser": "importai",
      "enabled": false,
      "_note": "5篇0阅读，VC方向不匹配",
      "check_interval_hours": 24,
      "limit": 5,
      "description": "VC and startup insights",
      "last_checked": "2026-03-02T03:16:25.800242+00:00"
    },
    {
      "id": "aitalks",
      "name": "AI Talks",
      "url": "https://aitalks.blog/feed/",
      "type": "newsletter_rss",
      "parser": "importai",
      "enabled": false,
      "_note": "7篇0阅读",
      "check_interval_hours": 24,
      "limit": 5,
      "description": "AI revolution navigation and weekly news",
      "last_checked": "2026-03-02T03:16:27.391858+00:00"
    },
    {
      "id": "kexuefm",
      "name": "科学空间 (Scientific Spaces)",
      "url": "https://kexue.fm/feed",
      "type": "blog_rss",
      "parser": "kexuefm",
      "enabled": true,
      "check_interval_hours": 168,
      "limit": 10,
      "last_checked": "2026-06-02T01:00:19.871397+00:00",
      "description": "苏剑林的AI/ML技术博客，MoE环游记系列"
    },
    {
      "id": "tylerfolkman",
      "name": "Tyler Folkman (AI Newsletter)",
      "url": "https://tylerfolkman.substack.com/",
      "type": "newsletter_rss",
      "parser": "tylerfolkman",
      "enabled": true,
      "check_interval_hours": 24,
      "limit": 3,
      "last_checked": "2026-06-02T01:00:19.871910+00:00"
    },
    {
      "id": "github_trending",
      "name": "GitHub Trending",
      "url": "https://github.com/trending",
      "type": "aggregator",
      "parser": "github_trending",
      "enabled": true,
      "check_interval_hours": 24,
      "limit": 10,
      "last_checked": "2026-06-02T01:00:19.871899+00:00"
    },
    {
      "id": "superhuman_ai",
      "name": "Superhuman AI",
      "url": "https://www.superhuman.ai",
      "type": "newsletter_sitemap",
      "parser": "superhuman_ai",
      "enabled": true,
      "check_interval_hours": 24,
      "limit": 3,
      "description": "AI工具与技术进展日报",
      "last_checked": "2026-06-02T01:00:19.871952+00:00"
    },
    {
      "id": "aibreakfast",
      "name": "AI Breakfast",
      "url": "https://aibreakfast.beehiiv.com",
      "type": "newsletter_sitemap",
      "parser": "aibreakfast",
      "enabled": true,
      "check_interval_hours": 168,
      "limit": 3,
      "description": "每周AI新闻与工具精选",
      "last_checked": "2026-06-02T01:00:19.872013+00:00"
    },
    {
      "id": "reddit_ai",
      "name": "Reddit AI (ML+LocalLLaMA+ClaudeAI+Anthropic)",
      "url": "https://www.reddit.com/r/MachineLearning+LocalLLaMA+ClaudeAI+Anthropic/.rss",
      "type": "aggregator_rss",
      "parser": "reddit_ai",
      "enabled": false,
      "_note": "用户要求禁用",
      "check_interval_hours": 24,
      "limit": 8,
      "description": "Reddit AI社区热帖：机器学习、本地LLM、Claude、Anthropic",
      "last_checked": "2026-04-18T08:45:50.442890+00:00"
    },
    {
      "id": "bestblogs",
      "name": "BestBlogs.dev",
      "url": "https://www.bestblogs.dev/articles",
      "type": "aggregator_html",
      "parser": "bestblogs",
      "enabled": true,
      "check_interval_hours": 24,
      "limit": 10,
      "description": "中文技术博客聚合：AI、编程、产品、商业精选",
      "last_checked": "2026-06-02T01:00:19.872041+00:00"
    },
    {
      "id": "simonwillison",
      "name": "Simon Willison's Weblog",
      "url": "https://simonwillison.net/atom/everything/",
      "type": "blog_rss",
      "parser": "importai",
      "enabled": true,
      "check_interval_hours": 24,
      "limit": 10,
      "description": "LLM/AI工具最活跃独立博主，Django联创，频繁评测Claude/GPT",
      "last_checked": "2026-06-02T01:00:19.871916+00:00"
    },
    {
      "id": "garymarcus",
      "name": "Gary Marcus (The Road to AI We Can Trust)",
      "url": "https://garymarcus.substack.com/feed",
      "type": "newsletter_rss",
      "parser": "importai",
      "enabled": true,
      "check_interval_hours": 168,
      "limit": 5,
      "description": "AI批评视角，提供平衡观点素材",
      "last_checked": "2026-06-02T01:00:19.871930+00:00"
    },
    {
      "id": "dwarkesh",
      "name": "Dwarkesh Patel",
      "url": "https://www.dwarkeshpatel.com/feed",
      "type": "blog_rss",
      "parser": "importai",
      "enabled": true,
      "check_interval_hours": 168,
      "limit": 5,
      "description": "深度访谈AI领袖和技术思想家",
      "last_checked": "2026-06-02T01:00:19.872026+00:00"
    },
    {
      "id": "gwern",
      "name": "Gwern Branwen",
      "url": "https://gwern.substack.com/feed",
      "type": "newsletter_rss",
      "parser": "importai",
      "enabled": true,
      "check_interval_hours": 168,
      "limit": 5,
      "description": "AI/未来主义/深度长文思考",
      "last_checked": "2026-06-02T01:00:19.871943+00:00"
    },
    {
      "id": "geoffreylitt",
      "name": "Geoffrey Litt",
      "url": "https://www.geoffreylitt.com/feed.xml",
      "type": "blog_rss",
      "parser": "importai",
      "enabled": true,
      "check_interval_hours": 168,
      "limit": 5,
      "description": "AI/HCI研究，工具设计思考",
      "last_checked": "2026-06-02T01:00:19.871983+00:00"
    },
    {
      "id": "experimental-history",
      "name": "Experimental History",
      "url": "https://www.experimental-history.com/feed",
      "type": "newsletter_rss",
      "parser": "importai",
      "enabled": true,
      "check_interval_hours": 168,
      "limit": 5,
      "description": "社会科学实验视角，独特洞察",
      "last_checked": "2026-06-02T01:00:19.871967+00:00"
    },
    {
      "id": "construction-physics",
      "name": "Construction Physics",
      "url": "https://www.construction-physics.com/feed",
      "type": "newsletter_rss",
      "parser": "importai",
      "enabled": true,
      "check_interval_hours": 168,
      "limit": 5,
      "description": "工程进步与科技发展分析",
      "last_checked": "2026-06-02T01:00:19.871980+00:00"
    },
    {
      "id": "antirez",
      "name": "Salvatore Sanfilippo (antirez)",
      "url": "http://antirez.com/rss",
      "type": "blog_rss",
      "parser": "importai",
      "enabled": true,
      "check_interval_hours": 168,
      "limit": 5,
      "description": "Redis作者博客，技术洞察",
      "last_checked": "2026-06-02T01:00:19.872039+00:00"
    },
    {
      "id": "theresanaiforthat",
      "name": "There's An AI For That",
      "url": "https://newsletter.theresanaiforthat.com",
      "type": "newsletter_sitemap",
      "parser": "beehiiv_sitemap",
      "enabled": true,
      "check_interval_hours": 168,
      "limit": 3,
      "description": "AI工具发现平台官方通讯，每周精选AI工具推荐",
      "last_checked": "2026-06-02T01:00:19.871986+00:00"
    }
  ]
}

{
  "_meta": {
    "version": "2.2",
    "updated_at": "2026-03-02",
    "updated_by": "基于全量阅读数据 + 第10周反馈优化（2500+篇行为分析）"
  },

  "user_profile": {
    "name": "Joe",
    "writing_style": "qiaomu",

    "primary_interests": [
      "AI Research Papers & Surveys",
      "AI Agents & MCP/Tool Use",
      "AI Reasoning & Thinking",
      "Interactive Technical Explainers",
      "Practical AI Tools & Reviews",
      "Show HN Projects",
      "Programming Craft & Engineering Philosophy",
      "Web Standards & Browser APIs",
      "Human Psychology & Social Insights"
    ],

    "secondary_interests": [
      "AI Video/Vision Generation",
      "Multimodal AI",
      "Diffusion Models",
      "SVG/Design Engineering",
      "Career & Founder Stories",
      "Open Source Culture",
      "Creative Tools",
      "Performance Engineering"
    ],

    "excluded_topics": [
      "DevOps Infrastructure",
      "Cloud Platform Admin",
      "Enterprise SaaS Marketing",
      "Crypto/Web3 Speculation",
      "Political News",
      "Sports",
      "AI Safety & Governance Policy",
      "AI Ethics & Regulation Debates",
      "Enterprise AI Strategy & Management",
      "Executive Briefings & Corporate AI Adoption",
      "Terminal Emulators & Shell Tools",
      "Effective Altruism",
      "VC/Investment Analysis"
    ]
  },

  "recommendation_preferences": {
    "min_articles_per_topic": 2,
    "max_recommendations": 10,

    "preferred_sources": [
      "Hacker News",
      "Hugging Face Papers",
      "NLP Newsletter (Elvis Saravia)",
      "Readwise Wise",
      "Import AI",
      "Interconnects (Nathan Lambert)",
      "科学空间 (Scientific Spaces)",
      "Ben's Bites",
      "TLDR AI",
      "Latent Space Podcast",
      "Paul Graham Essays",
      "Wait But Why",
      "Why Try AI",
      "One Useful Thing (Ethan Mollick)"
    ],

    "scoring_weights": {
      "interest_match": 0.45,
      "depth_potential": 0.25,
      "timeliness": 0.15,
      "source_reputation": 0.15
    }
  },

  "content_filters": {
    "high_priority_keywords": [
      "Video Generation", "Video Understanding", "4D", "3D Scene",
      "Vision Language Model", "Multimodal", "Visual",
      "World Model", "Agent", "Agentic", "Autonomous",
      "Reasoning", "Thinking", "Test-Time", "Chain of Thought",
      "Diffusion", "Transformer", "Attention",
      "Show HN", "I built", "I made",
      "Interactive", "Explainer", "Playground", "Visualiz",
      "Top AI Papers", "Survey", "Benchmark",
      "SVG", "Design Tool", "Arrow", "Diagram",
      "Compiler", "Algorithm", "Data Structure",
      "MCP", "Tool Use", "Function Calling", "WebMCP",
      "AGENTS.md", "CLAUDE.md", "Context Protocol",
      "from scratch", "in 200 lines", "minimal implementation"
    ],

    "medium_priority_keywords": [
      "LLM", "GPT", "Claude", "ChatGPT", "OpenAI", "Anthropic",
      "Training", "Fine-tuning", "RLHF", "RLVR", "Reinforcement",
      "Image Generation", "Audio", "TTS", "Speech",
      "Embedding", "RAG", "Context Window",
      "Open Source", "GitHub",
      "Founder", "Startup", "CEO", "Career",
      "Linux", "Kernel", "Systems Programming",
      "Performance", "Optimization", "Fast", "Speed",
      "Psychology", "Social", "Behavior", "Human"
    ],

    "exclude_keywords": [
      "Kubernetes", "Docker Swarm", "Terraform", "Ansible",
      "Enterprise", "B2B SaaS Marketing",
      "Crypto", "Bitcoin", "Blockchain", "NFT", "Web3",
      "Political", "Election", "Government Policy",
      "AI Safety", "AI Governance", "AI Regulation", "AI Ethics",
      "Situational Awareness", "Effective Altruism",
      "Executive Briefing", "Corporate AI Adoption",
      "Terminal Emulator", "Ghostty", "Wezterm", "Alacritty",
      "VC Funding", "Series A", "Series B", "Fundraising Round"
    ],

    "source_specific_filters": {
      "Hugging Face Papers": {
        "prefer": ["Video", "Vision", "Agent", "World Model", "Reasoning", "Diffusion", "Tool Use"],
        "min_votes": 20
      },
      "Hacker News": {
        "prefer": ["Show HN", "AI", "Interactive", "Programming", "Performance", "MCP", "from scratch"],
        "avoid": ["political", "sports", "celebrity", "terminal emulator", "ghostty"]
      }
    }
  },

  "writing_goals": {
    "article_types": [
      "深度技术解读（2000-3000字）- 论文/系统设计",
      "工具体验分享（1000-1500字）- Show HN 项目",
      "认知升级（1500-2000字）- 方法论/思维方式",
      "行业观察（1000-1500字）- 趋势/人物"
    ],
    "target_audience": "对AI和编程感兴趣、有一定技术背景的读者",
    "tone": "conversational, insightful, technically grounded"
  },

  "discovered_patterns": {
    "reading_behavior": {
      "total_articles": 2500,
      "read_rate": "10%",
      "star_rate_when_read": "88% (HN), 100% (HF)",
      "primary_reading_sources": "HN (172 read) + HF (59 read) = 96% of all reads",
      "never_read_sources": "KDnuggets, HackerNoon (all 3), Neuron Daily, Rundown AI, 多个substack"
    },
    "week10_feedback": {
      "excited_about": ["AI Papers周报", "MicroGPT交互讲解", "QuiverAI Arrow SVG", "WebMCP标准"],
      "followed_with_interest": ["MCP vs CLI辩论", "AGENTS.md研究", "PAHF个性化Agent"],
      "not_interested": ["Cognitive Revolution (AI policy)", "Nate's Newsletter (enterprise)", "Ghostty (terminal)"]
    },
    "top_hf_categories": [
      "视觉/视频生成 (29%)",
      "Agent/世界模型 (15%)",
      "扩散/生成 (14%)",
      "推理/思考 (14%)"
    ],
    "top_hn_categories": [
      "技术深度文章 (62%)",
      "AI相关 (13%)",
      "Linux/系统 (7%)",
      "Show HN 项目 (7%)"
    ],
    "source_preference": {
      "Hacker News": "71% of starred",
      "Hugging Face Papers": "28% of starred"
    }
  }
}

MIT License

Copyright (c) 2026 joeseesun

Permission is hereby granted, free of charge, to any person obtaining a copy
of this software and associated documentation files (the "Software"), to deal
in the Software without restriction, including without limitation the rights
to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
copies of the Software, and to permit persons to whom the Software is
furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in all
copies or substantial portions of the Software.

THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
SOFTWARE.

{
  "name": "qiaomu-ai-radar",
  "version": "1.0",
  "description": "每日选题助手 - 监控多个内容源，自动采集最新文章并推荐值得创作的话题。支持James Clear、Ben's Bites、Wait But Why、Hacker News、Product Hunt等。",
  "author": "joe",
  "triggers": [
    "检查今天的更新",
    "抓取最新文章",
    "更新内容库",
    "今天值得写什么",
    "推荐选题",
    "有什么好话题"
  ],
  "features": [
    "自动监控5+内容源",
    "URL去重，只关注新内容",
    "macOS定时任务支持（每天9am）",
    "智能选题推荐，结合乔木写作风格",
    "多来源热度分析"
  ],
  "created": "2025-12-24"
}

#!/bin/bash
# ============================================================
# 每日自动抓取 + macOS通知 + Obsidian集成
# v2.8 - 自动写入当日日记
# ============================================================

SKILL_DIR="$HOME/.agents/skills/qiaomu-ai-radar"
LOG_FILE="$SKILL_DIR/data/auto_run.log"
OUTPUT_DIR="$HOME/乔木新知识库/20-29 学习/21 素材索引"
DIARY_DIR="$HOME/乔木新知识库/30-39 成长/35 日记"
export PATH="$HOME/.local/node/bin:/opt/homebrew/bin:/usr/local/bin:/usr/bin:/bin:$PATH"

# 记录开始时间
echo "========================================" >> "$LOG_FILE"
echo "自动运行开始: $(date '+%Y-%m-%d %H:%M:%S')" >> "$LOG_FILE"

# 切换到scripts目录
cd "$SKILL_DIR/scripts" || exit 1

# 运行统一入口：抓取、翻译、飞书同步、验证、打开 Obsidian
python3 run_radar.py --mode fetch >> "$LOG_FILE" 2>&1
EXIT_CODE=$?

# 计算当前周数，构建周刊文件路径
YEAR=$(date '+%Y')
WEEK=$(date '+%V' | sed 's/^0//')
WEEKLY_FILE="$OUTPUT_DIR/${YEAR}-第${WEEK}周.md"

# 检查是否成功
if [ $EXIT_CODE -eq 0 ]; then
    # 统计新文章数量
    NEW_COUNT=$(tail -20 "$LOG_FILE" | grep "总计发现" | tail -1 | grep -oE "[0-9]+ 篇" | grep -oE "[0-9]+")

    if [ -z "$NEW_COUNT" ]; then
        NEW_COUNT=0
    fi

    if [ "$NEW_COUNT" -gt 0 ]; then
        # 快速通知（通知中心）
        osascript -e "display notification \"发现 $NEW_COUNT 篇新文章，已同步到素材库\" with title \"📚 每日选题助手\" sound name \"Glass\""

        # 可交互对话框（带打开按钮）
        if [ -n "$WEEKLY_FILE" ] && [ -f "$WEEKLY_FILE" ]; then
            # URL编码文件路径
            ENCODED_PATH=$(python3 - "$WEEKLY_FILE" << 'PYEOF'
import sys
from pathlib import Path
from urllib.parse import quote

path = Path(sys.argv[1]).with_suffix("")
vault = Path.home() / "乔木新知识库"
try:
    rel = path.relative_to(vault)
except ValueError:
    rel = path
print(f"obsidian://open?vault={quote(vault.name, safe='')}&file={quote(str(rel), safe='')}")
PYEOF
)
            osascript <<EOF
tell application "System Events"
    activate
    set userChoice to button returned of (display dialog "发现 $NEW_COUNT 篇新文章！

📊 已更新本周索引
📂 $(basename "$WEEKLY_FILE")" buttons {"稍后查看", "打开 Obsidian"} default button "打开 Obsidian" with title "📚 每日选题助手" with icon note)

    if userChoice is "打开 Obsidian" then
        do shell script "open '$ENCODED_PATH'"
    end if
end tell
EOF
        fi

        echo "✅ 成功: 发现 $NEW_COUNT 篇新文章" >> "$LOG_FILE"
    else
        # 无新文章通知（对话框 + 通知中心）
        osascript -e "display notification \"今日暂无新内容\" with title \"📚 每日选题助手\""
        osascript -e "display dialog \"今日暂无新内容\\n\\n所有信息源均已检查\" with title \"📚 每日选题助手\" buttons {\"知道了\"} default button \"知道了\" giving up after 10"
        echo "ℹ️  今日无新文章" >> "$LOG_FILE"
    fi

    # ── 写入当日日记 ──────────────────────────────────────────
    TODAY=$(date '+%Y-%m-%d')
    WEEKDAY=$(LC_TIME=zh_CN.UTF-8 date '+%A' 2>/dev/null || date '+%A')
    DIARY_FILE="$DIARY_DIR/${TODAY}.md"
    WEEK=$(date '+%V' | sed 's/^0//')
    YEAR=$(date '+%Y')
    WEEKLY_LINK="[[../../20-29 学习/21 素材索引/${YEAR}-第${WEEK}周]]"
    SECTION="## 今日阅读素材

> 由 Topic Selector 自动生成，点击查看 → ${WEEKLY_LINK}
"

    if [ ! -f "$DIARY_FILE" ]; then
        # 日记不存在：用模板创建
        cat > "$DIARY_FILE" << EOF
# ${TODAY} ${WEEKDAY}

${SECTION}
---

## 随手记



---

## 晚间复盘

EOF
        echo "📔 创建当日日记: $DIARY_FILE" >> "$LOG_FILE"
    else
        # 日记已存在：检查是否已有「今日阅读素材」节
        if ! grep -q "今日阅读素材" "$DIARY_FILE"; then
            # 没有这一节，用 Python 插到标题行之后
            python3 - "$DIARY_FILE" "$WEEKLY_LINK" << 'PYEOF'
import sys

diary_path = sys.argv[1]
weekly_link = sys.argv[2]

with open(diary_path, 'r', encoding='utf-8') as f:
    lines = f.readlines()

section = f"\n## 今日阅读素材\n\n> 由 Topic Selector 自动生成，点击查看 → {weekly_link}\n\n"

# 在第一行（标题）之后插入
if lines:
    lines.insert(1, section)
else:
    lines = [section]

with open(diary_path, 'w', encoding='utf-8') as f:
    f.writelines(lines)

print("done")
PYEOF
            echo "📔 已在日记中插入今日阅读素材" >> "$LOG_FILE"
        else
            echo "📔 日记已有今日阅读素材，跳过" >> "$LOG_FILE"
        fi
    fi
    # ─────────────────────────────────────────────────────────
else
    # 发送失败通知
    osascript -e "display notification \"抓取过程出现错误，请查看日志\" with title \"⚠️  每日选题助手\" sound name \"Basso\""
    echo "❌ 失败: 退出码 $EXIT_CODE" >> "$LOG_FILE"
fi

echo "结束时间: $(date '+%Y-%m-%d %H:%M:%S')" >> "$LOG_FILE"
echo "" >> "$LOG_FILE"

#!/usr/bin/env python3
"""
Translate untranslated weekly index titles.

The weekly file is the source of truth for reading. This script only touches
title lines in the current checkbox format:

    - [ ] **标题**: English Title

Already bilingual lines containing " / " are left unchanged.
"""

import argparse
import json
import os
import re
import sys
import time
from datetime import datetime
from pathlib import Path
from typing import Dict, List, Optional
from urllib import request, error

SCRIPT_DIR = Path(__file__).parent
sys.path.insert(0, str(SCRIPT_DIR / "shared"))

from path_utils import get_weekly_index_path  # noqa: E402

TITLE_RE = re.compile(r"^(?P<prefix>\s*- \[[ xX]\] \*\*标题\*\*: )(?P<title>.+?)\s*$")
CJK_RE = re.compile(r"[\u3400-\u9fff]")
REPO_SLUG_RE = re.compile(r"^[A-Za-z0-9_.-]+/[A-Za-z0-9_.-]+$")
URL_RE = re.compile(r"^https?://", re.IGNORECASE)

TECH_TERMS = {
    "AGI",
    "AI",
    "API",
    "Claude",
    "DeepSeek",
    "GPT",
    "GPU",
    "LLM",
    "MCP",
    "OpenAI",
    "SDK",
    "Transformer",
}


def current_weekly_file() -> Path:
    now = datetime.now()
    return get_weekly_index_path(now.year, now.isocalendar()[1])


def is_untranslated_title(title: str) -> bool:
    if " / " in title or CJK_RE.search(title):
        return False
    if REPO_SLUG_RE.match(title) or URL_RE.match(title):
        return False
    return True


def load_titles(content: str) -> List[Dict[str, object]]:
    titles = []
    for index, line in enumerate(content.splitlines()):
        match = TITLE_RE.match(line)
        if not match:
            continue
        title = match.group("title").strip()
        titles.append(
            {
                "line": index,
                "prefix": match.group("prefix"),
                "title": title,
                "untranslated": is_untranslated_title(title),
            }
        )
    return titles


def translate_with_api(title: str, provider: str, timeout: int = 45) -> Optional[str]:
    if provider == "none":
        return None

    if provider == "deepseek":
        api_key = os.environ.get("DEEPSEEK_API_KEY")
        base_url = os.environ.get("DEEPSEEK_BASE_URL", "https://api.deepseek.com")
        model = os.environ.get("QIAOMU_TRANSLATE_MODEL", "deepseek-chat")
        url = f"{base_url.rstrip('/')}/chat/completions"
    else:
        api_key = os.environ.get("OPENAI_API_KEY")
        base_url = os.environ.get("OPENAI_BASE_URL", "https://api.openai.com/v1")
        model = os.environ.get("QIAOMU_TRANSLATE_MODEL", "gpt-5-mini")
        url = f"{base_url.rstrip('/')}/chat/completions"

    if not api_key:
        return None

    system = (
        "Translate article titles into concise natural Chinese. Preserve these "
        f"technical/proper terms when appropriate: {', '.join(sorted(TECH_TERMS))}. "
        "Return only the Chinese title, no quotes, no explanation."
    )
    payload = {
        "model": model,
        "messages": [
            {"role": "system", "content": system},
            {"role": "user", "content": title},
        ],
        "temperature": 0.2,
    }
    data = json.dumps(payload).encode("utf-8")
    req = request.Request(
        url,
        data=data,
        headers={
            "Authorization": f"Bearer {api_key}",
            "Content-Type": "application/json",
        },
        method="POST",
    )

    try:
        with request.urlopen(req, timeout=timeout) as resp:
            body = json.loads(resp.read().decode("utf-8"))
    except (error.HTTPError, error.URLError, TimeoutError, json.JSONDecodeError) as exc:
        raise RuntimeError(f"{provider} translation failed for {title!r}: {exc}") from exc

    choices = body.get("choices") or []
    if not choices:
        raise RuntimeError(f"{provider} translation returned no choices for {title!r}")
    text = choices[0].get("message", {}).get("content", "").strip()
    return re.sub(r"^['\"]|['\"]$", "", text).strip() or None


def choose_provider(provider: str) -> str:
    if provider != "auto":
        return provider
    if os.environ.get("DEEPSEEK_API_KEY"):
        return "deepseek"
    if os.environ.get("OPENAI_API_KEY"):
        return "openai"
    return "none"


def translate_weekly(
    weekly_file: Path,
    dry_run: bool = False,
    provider: str = "auto",
    limit: Optional[int] = None,
) -> Dict[str, object]:
    weekly_file = weekly_file.expanduser()
    result: Dict[str, object] = {
        "ok": False,
        "weekly_file": str(weekly_file),
        "provider": choose_provider(provider),
        "dry_run": dry_run,
        "title_count": 0,
        "untranslated_before": 0,
        "translated_count": 0,
        "untranslated_after": 0,
        "errors": [],
    }

    if not weekly_file.exists():
        result["errors"] = [f"weekly file not found: {weekly_file}"]
        return result

    content = weekly_file.read_text(encoding="utf-8")
    lines = content.splitlines()
    titles = load_titles(content)
    untranslated = [item for item in titles if item["untranslated"]]
    if limit is not None:
        untranslated = untranslated[:limit]

    result["title_count"] = len(titles)
    result["untranslated_before"] = len([item for item in titles if item["untranslated"]])

    provider_name = result["provider"]
    for item in untranslated:
        title = str(item["title"])
        if provider_name == "none":
            continue
        try:
            translated = translate_with_api(title, str(provider_name))
        except RuntimeError as exc:
            result["errors"].append(str(exc))
            continue
        if not translated:
            continue
        new_title = f"{title} / {translated}"
        lines[int(item["line"])] = f"{item['prefix']}{new_title}"
        result["translated_count"] = int(result["translated_count"]) + 1
        time.sleep(0.2)

    new_content = "\n".join(lines)
    if content.endswith("\n"):
        new_content += "\n"

    after_titles = load_titles(new_content)
    result["untranslated_after"] = len([item for item in after_titles if item["untranslated"]])

    if not dry_run and new_content != content:
        weekly_file.write_text(new_content, encoding="utf-8")

    result["ok"] = not result["errors"]
    if provider_name == "none" and result["untranslated_before"]:
        result["ok"] = False
        result["errors"].append(
            "No translation provider configured. Set DEEPSEEK_API_KEY or OPENAI_API_KEY."
        )
    return result


def main() -> int:
    parser = argparse.ArgumentParser(description="翻译乔木 AI 雷达本周周刊标题")
    parser.add_argument("weekly_file", nargs="?", help="周刊 Markdown 文件路径；不传则使用本周文件")
    parser.add_argument("--dry-run", action="store_true", help="只检查/调用翻译，不写回文件")
    parser.add_argument("--json", action="store_true", help="输出 JSON")
    parser.add_argument(
        "--provider",
        choices=["auto", "deepseek", "openai", "none"],
        default="auto",
        help="翻译提供方，默认自动按环境变量选择",
    )
    parser.add_argument("--limit", type=int, help="最多处理多少个未翻译标题")
    args = parser.parse_args()

    weekly_file = Path(args.weekly_file).expanduser() if args.weekly_file else current_weekly_file()
    result = translate_weekly(
        weekly_file=weekly_file,
        dry_run=args.dry_run,
        provider=args.provider,
        limit=args.limit,
    )

    if args.json:
        print(json.dumps(result, ensure_ascii=False, indent=2))
    else:
        print(f"周刊: {result['weekly_file']}")
        print(f"翻译提供方: {result['provider']}")
        print(f"标题数: {result['title_count']}")
        print(f"翻译前未翻译: {result['untranslated_before']}")
        print(f"本次实际翻译: {result['translated_count']}")
        print(f"翻译后未翻译: {result['untranslated_after']}")
        for message in result["errors"]:
            print(f"ERROR: {message}", file=sys.stderr)

    return 0 if result["ok"] else 1


if __name__ == "__main__":
    raise SystemExit(main())