[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"project-78194":3},{"id":4,"name":5,"fullName":6,"owner":7,"repo":5,"description":8,"homepage":9,"htmlUrl":9,"language":9,"languages":9,"totalLinesOfCode":9,"stars":10,"forks":11,"watchers":12,"openIssues":13,"contributorsCount":14,"subscribersCount":14,"size":14,"stars1d":14,"stars7d":12,"stars30d":15,"stars90d":14,"forks30d":14,"starsTrendScore":14,"compositeScore":16,"rankGlobal":9,"rankLanguage":9,"license":17,"archived":18,"fork":18,"defaultBranch":19,"hasWiki":20,"hasPages":18,"topics":21,"createdAt":9,"pushedAt":9,"updatedAt":30,"readmeContent":31,"aiSummary":32,"trendingCount":14,"starSnapshotCount":14,"syncStatus":33,"lastSyncTime":34,"discoverSource":35},78194,"POHA","jigripokri\u002FPOHA","jigripokri","Personal Overnight Helper Agent. Runs while you sleep. Serves up a morning brief before your alarm.",null,120,41,1,3,0,18,47.17,"MIT License",false,"master",true,[22,23,24,25,26,27,28,29],"ai-assistant","claude","claude-code","morning-brief","personal-assistant","poha","productivity","second-brain","2026-06-12 04:01:23","# POHA\n\n**Personal Overnight Helper Agent. Runs while you sleep. Serves up a morning brief before your alarm. Forks in 30 minutes. Runs on your existing Claude subscription.**\n\nBuilt an AI that runs overnight and serves up a morning brief. Called it POHA, because of course.\n\nThe pun completes itself: POHA is the breakfast you wake up to. This agent literally runs overnight and delivers a morning brief before your alarm. The dish IS the deliverable.\n\n---\n\n## What POHA is (and isn't)\n\nMost \"AI assistants\" are reactive. You open them, you type, they answer, you close them, nothing persists. That's a tool, not a brain extension.\n\nPOHA has three properties:\n\n1. **It observes.** Reads your inbound — WhatsApp, Gmail, Calendar, Tasks, SMS — without you having to ask.\n2. **It remembers.** Persistent state across sessions, in plain markdown you can edit by hand.\n3. **It acts on a schedule.** Runs while you sleep. The morning briefing is on your phone before your alarm.\n\nThis repo wires all three.\n\n---\n\n## What you get on day one\n\nSeven scheduled tasks already wired up:\n\n- **5am daily — Morning briefing.** Scans your WhatsApp groups + DMs, Gmail, Calendar, and tasks. Flags commitments, unanswered threads, and what's on your plate today. Emails you the result with one-tap acknowledgment links for every action item.\n- **9pm daily — Evening reflection.** Catches plans confirmed after the morning pass, updates memory.\n- **Sunday 9am — Weekly review.** Prunes stale memory (3-week rule), refreshes priorities, reviews relationship health.\n- **7am daily — Commitment → Calendar.** Turns every \"let's grab coffee Saturday\" into a real calendar event.\n- **Friday noon — GSD agenda.** Builds a punch list from your tasks + open commitments for a Friday focus block.\n- **Friday noon — Weekend Fun Finder.** Searches local events, scored against your household's preferences.\n- **28th monthly — Financial sweep.** Spending autopsy, net worth update, subscription audit, credit card optimization.\n\nFive on-demand skills:\n\n- `\u002Fdraft` — Writes outbound messages in your voice. Studies the last 10–20 messages in any thread before drafting.\n- `\u002Freply` — Three reply options + a recommendation for any message thread.\n- `\u002Fwildcard` — One contrarian, non-obvious insight on whatever's on your mind.\n- `\u002Fgsd` — Prioritized punch list, right now.\n- `\u002Froast` — Group-chat roast on demand. Tune it to your circle's tone.\n\n---\n\n## The three architectural unlocks\n\nMost personal-AI projects fail in predictable ways. Here's what we did differently.\n\n**1. Persistent memory in plain markdown.** Seven files. Claude reads them at the start of every task, updates them at the end. No vector DB. No RAG. No fine-tuning. Just files. The reason this works: markdown is human-editable, version-controllable, and grep-able. You can correct POHA's memory by opening a file in any editor. You own the state. (See [`docs\u002Fmemory-system.md`](.\u002Fdocs\u002Fmemory-system.md).)\n\n**2. The acknowledgment loop.** This is the hardest problem in personal AI: how do you close the loop on actions that happen offline? You called your mom. There's no API event. The agent has no way to know.\n\nSolution: every briefing email includes one-tap `mailto:` links next to action items. Tapping one fires a self-email with a coded subject — `EXO_DONE::call-mom` — that the next briefing parses. Match the slug, mark the commitment done, archive the signal. (See [`docs\u002Facknowledgment-system.md`](.\u002Fdocs\u002Facknowledgment-system.md).)\n\n**3. Tier-weighted attention.** Not all messages are equal. The morning sweep applies five tiers (Family → Extended family → Close friends → Network → Other), so a \"hey when are you free?\" from your partner surfaces before a meme from a group chat. The tier system is in `CLAUDE.md` — change it to match your life. (See [`docs\u002Fcustomization.md`](.\u002Fdocs\u002Fcustomization.md).)\n\n---\n\n## Architecture\n\nThree layers. Full breakdown in [`ARCHITECTURE.md`](.\u002FARCHITECTURE.md).\n\n```\n┌──────────────────────────────────────────────────────────┐\n│ Layer 1 — The Briefing Engine (scheduled tasks)           │\n│   morning · evening · weekly · daily commitment sweep ·   │\n│   GSD agenda · weekend finder · monthly finance           │\n├──────────────────────────────────────────────────────────┤\n│ Layer 2 — The Memory System (seven markdown files)        │\n│   people · commitments · life · insights ·                │\n│   finances · health · carry-on (private, gated)           │\n├──────────────────────────────────────────────────────────┤\n│ Layer 3 — Skills (on-demand commands)                     │\n│   \u002Fdraft · \u002Freply · \u002Fwildcard · \u002Fgsd · \u002Froast             │\n└──────────────────────────────────────────────────────────┘\n         ▲\n         │ reads\u002Fwrites via MCP connectors\n         ▼\n┌──────────────────────────────────────────────────────────┐\n│ WhatsApp · Gmail · Google Calendar · Google Tasks ·       │\n│ Android Messages · Fitbit · Monarch Money                 │\n└──────────────────────────────────────────────────────────┘\n```\n\n---\n\n## Quickstart\n\nFull walkthrough in [`SETUP.md`](.\u002FSETUP.md). The 30-second version:\n\n1. Install [Claude Desktop](https:\u002F\u002Fclaude.ai\u002Fdownload) with Cowork mode enabled.\n2. Clone this repo somewhere stable on your machine.\n3. Mount the folder in Cowork.\n4. Open [`CLAUDE.md`](.\u002FCLAUDE.md) and fill in the placeholders (`{{USER_NAME}}`, `{{PARTNER_NAME}}`, your tier system, etc.).\n5. Connect the MCPs you want — WhatsApp, Gmail, Calendar are the high-leverage trio.\n6. Tell Claude: *\"Read CLAUDE.md and onboard me. Walk me through filling in the memory files.\"*\n7. Once memory is bootstrapped, tell Claude: *\"Set up the seven scheduled tasks per BACKLOG.md.\"*\n\nThe first morning brief lands at 5am the next day. Hot and ready, like the dish.\n\n---\n\n## Customize\n\nThree knobs to turn:\n\n**Persona.** Edit the `Persona` section of `CLAUDE.md`. The default is \"direct, warm, slightly cheeky chief of staff who tells you what you need to hear, not what you want to hear.\" If you want softer or more sarcastic, change it there. Every output inherits.\n\n**Tier system.** The default tiers are placeholder categories (Family \u002F Meaningful work \u002F Health). Yours might be Cofounders \u002F Investors \u002F Side project. Whatever they are, write them down — POHA filters everything through this lens.\n\n**Scheduled tasks.** Each briefing in [`briefings\u002F`](.\u002Fbriefings) is a standalone prompt. Edit them, remove the ones you don't want, add new ones for your situation. The founder version probably wants a daily \"what moved the needle?\" reflection. The new-parent version probably wants a school-pickup buffer alert.\n\nSee [`docs\u002Fcustomization.md`](.\u002Fdocs\u002Fcustomization.md).\n\n---\n\n## What it costs\n\n- **Claude Pro or Max subscription**: already what you're paying. Max recommended if you're running all seven scheduled tasks against a busy life.\n- **MCP connectors**: free for personal accounts (Gmail, Calendar, WhatsApp, Fitbit). Monarch Money is $15\u002Fmo if you want financial sweeps.\n- **Your time**: ~3 hours upfront to bootstrap memory honestly. ~5 min\u002Fweek to review what POHA surfaces. The 3 hours pay back inside a month.\n\n---\n\n## Who this is for\n\nYou will get value out of this if you:\n\n- Run a complicated life (family + work + side projects + community + investing) and lose threads regularly.\n- Already use Claude a lot and feel the pain of starting from scratch every conversation.\n- Are willing to write down what actually matters to you in markdown.\n\nYou will not get value if you:\n\n- Want a chatbot you talk to. POHA runs in the background; you barely interact with it.\n- Are uncomfortable letting an LLM read your messages and email. It does. It has to.\n- Want zero configuration. The configuration *is* the product — your context is what makes it work.\n\n---\n\n## Privacy & safety\n\n- All memory lives on your machine. Nothing leaves your laptop except Claude API calls and the briefing emails you send to yourself.\n- The `memory\u002Fprivate\u002Fcarry-on.md` file is *strictly gated* — never read automatically, never referenced in any output, only loaded when you explicitly invoke it. Use it for anything sensitive.\n- `.gitignore` excludes `memory\u002F` from version control by default. Your second brain stays out of GitHub.\n- Read [`docs\u002Fprivacy.md`](.\u002Fdocs\u002Fprivacy.md) before connecting anything financial or medical.\n\n---\n\n## Repo map\n\n```\npoha\u002F\n├── README.md                  ← you are here\n├── ARCHITECTURE.md            ← the three-layer system, in depth\n├── SETUP.md                   ← step-by-step onboarding\n├── BACKLOG.md                 ← what to build next (yours to fork)\n├── CLAUDE.md                  ← the persona + ruleset (fill in placeholders)\n├── memory\u002F                    ← your second brain (gitignored)\n│   ├── people.md.template\n│   ├── commitments.md.template\n│   ├── life.md.template\n│   ├── insights.md.template\n│   ├── finances.md.template\n│   ├── health.md.template\n│   └── private\u002Fcarry-on.md.template\n├── briefings\u002F                 ← the seven scheduled task prompts\n├── skills\u002F                    ← \u002Fdraft, \u002Freply, \u002Fwildcard, \u002Fgsd, \u002Froast\n├── docs\u002F                      ← deep dives on tricky bits\n└── examples\u002F                  ← a sanitized example morning briefing\n```\n\n---\n\n## Contributing\n\nIf this is useful, star the repo — it helps others find it. If you've extended it in interesting ways (new briefing types, new connectors, new skills), open a PR. I'd especially love to see versions tuned for founders, students, parents of young kids, and retirees.\n\n## License\n\nMIT. Take it, fork it, sell it, whatever. Just don't claim you invented the acknowledgment loop without crediting it — that one took a while to figure out.\n","POHA是一款个人夜间助手代理，它在你睡觉时运行，并在闹钟响起前为你提供早晨简报。该项目的核心功能包括自动读取你的WhatsApp、Gmail、日历和任务等信息，无需手动请求；保持会话间的状态持久化，以易于编辑的Markdown格式存储；以及按照设定的时间表执行任务。特别适合需要提高日常效率、希望早晨快速了解当天重要事项的用户使用。POHA通过简单的文件系统管理记忆，避免了复杂的数据库或微调需求，使得用户能够轻松地查看和修改AI的记忆内容。",2,"2026-06-11 03:56:33","CREATED_QUERY"]