Skip to content
Back to search

daily-briefing-agent

MarketplaceWorks with Claude CodeGitHub

A morning brief, an evening review and a signals digest, with every failed step printed at the top.

Safety scan

A static text scan, not a guarantee. Select the badge to see what was scanned and every finding.

Checked today
Badge for your README
findsafeskills safety rating

Markdown

[![findsafeskills safety rating](/api/listings/mavrin-xv--daily-briefing-agent/badge.svg)](/listings/mavrin-xv--daily-briefing-agent)

HTML

<a href="/listings/mavrin-xv--daily-briefing-agent"><img src="/api/listings/mavrin-xv--daily-briefing-agent/badge.svg" alt="findsafeskills safety rating" /></a>

Install

Install instructions from the README

As written by the repo's author. findsafeskills has not verified these steps.

1. Install and create a workspace:

   ./install.sh --workspace ~/briefing
   cd ~/briefing

2. Fill in `notes/priorities.md` (goals, open decisions, topics you track) and put your feeds in `sources.json`. 3. Ask for a brief in Claude Code, inside the workspace:

   > morning brief

4. When it reads right, schedule it (see `skills/brief-ticker/references/scheduling.md`) and set `TELEGRAM_BOT_TOKEN` and `TELEGRAM_CHAT_ID`.

Try it offline first, with the invented coffee roaster in `examples/`:

cd examples
python3 ../skills/brief-ticker/scripts/ticker.py --config ticker.demo.json

Sample summary (the full brief is in `examples/sample-morning-brief.md`):

Morning brief 2026-03-02: 1 step(s) need attention

[OK] signals: done
[OK] morning-brief: done
[FAIL] crm-export: capped (CRM API returned 401: token expired)
[..] evening-review: pending

--- Brief ---
# Morning brief, Mon 2 Mar 2026
System: 1 alert. Step crm-export failed (CRM token expired). Pipeline numbers below are from Friday.
**#1 today:** Call Northwind Bakery before 11:00. Open 4 days; today you either call or drop the lead.

What was scanned

The files the safety scan read, as written by the repository's author and shown as plain text. Lines with a finding are marked.

.claude-plugin/marketplace.json · Manifest · 15 lines
  1. {
  2. "name": "daily-briefing-agent",
  3. "plugins": [
  4. {
  5. "name": "daily-briefing-agent",
  6. "source": "./",
  7. "description": "A chief-of-staff agent: a two-minute morning brief, an evening review that forces decisions on stuck tasks, a deduped signals digest from feeds you pick, and one scheduled Telegram summary that lists every failed step instead of hiding it.",
  8. "version": "0.1.0"
  9. }
  10. ],
  11. "owner": {
  12. "name": "Artem Mavrin",
  13. "url": "https://github.com/mavrin-xv"
  14. }
  15. }
README.md · README · 145 lines
  1. ![daily-briefing-agent](assets/banner.png)
  2. # daily-briefing-agent
  3. **A chief-of-staff agent for solo founders and operators: a two-minute morning brief, an evening review, a signals digest from feeds you pick, and one Telegram message a day.**
  4. Your daily AI brief stops working quietly, and you keep reading confident summaries built on two-day-old data. This pack prints every failed step at the top of the message.
  5. ![License: MIT](https://img.shields.io/badge/license-MIT-blue) ![Works with Claude Code](https://img.shields.io/badge/works%20with-Claude%20Code-orange) ![Python 3.8+ stdlib](https://img.shields.io/badge/python-3.8%2B%20stdlib-green)
  6. ```bash
  7. git clone https://github.com/mavrin-xv/daily-briefing-agent && cd daily-briefing-agent
  8. ./install.sh --workspace ~/briefing # skills + a ready briefing folder
  9. ```
  10. <!-- install-notes:start -->
  11. Other ways in: as a Claude Code plugin, run `/plugin marketplace add mavrin-xv/daily-briefing-agent` and then `/plugin install daily-briefing-agent@daily-briefing-agent`. For Codex or another agent that reads `SKILL.md`, copy the folders under `skills/` into its skills directory. On Windows, run `install.sh` from Git Bash or WSL and type `py -3` where these docs say `python3`.
  12. <!-- install-notes:end -->
  13. ## What it solves
  14. | Pain | What the pack does |
  15. |---|---|
  16. | Morning starts in five apps and the inbox decides your day | `morning-brief`: one page, health first, then the **#1 action**, then decisions waiting on you |
  17. | The same task slides for a week and nobody says so | `evening-review` finds items open 3+ days and forces one verdict: first, date, hand off or drop |
  18. | News digests are forty links, half of them the same story | `signals-digest`: a script fetches, dedupes across sources and remembers what you saw; the model keeps only items that touch your goals, each with an action |
  19. | A feed dies and you notice a month later | Every source is reported `ok`, `empty` or `failed` in every digest |
  20. | Four cron jobs, one crashes, the others keep sending | `brief-ticker`: one scheduler entry, catch-up after sleep, retries with a cap, one summary that lists every step as `[OK]`, `[FAIL]` or `[MISS]` |
  21. | "Brief me on this client" means ten minutes of scrolling notes | `situation-brief`: 15 lines, where it stands, open questions, risks, next move, with sources |
  22. ## What's inside
  23. | Skill | What it does | Say |
  24. |---|---|---|
  25. | [morning-brief](skills/morning-brief/SKILL.md) | Health check, #1 action, today's top 3, waiting decisions, scored signals, what was not checked | "morning brief", "what's my day" |
  26. | [evening-review](skills/evening-review/SKILL.md) | Done vs planned, forced verdicts on stuck items, tomorrow's #1 written into tomorrow's note; weekly mode | "evening review", "weekly review" |
  27. | [signals-digest](skills/signals-digest/SKILL.md) | RSS, Atom, Reddit, YouTube, GitHub releases, newsletters: fetch, dedupe, seen-state, score against your goals | "signals", "what's new in my niche" |
  28. | [situation-brief](skills/situation-brief/SKILL.md) | One subject before a call: state, last change, open questions, risks, next move | "brief me on Northwind" |
  29. | [brief-ticker](skills/brief-ticker/SKILL.md) | The scheduler: one tick every 30 min, lock, caps, catch-up, one summary, zero-token health check, Telegram | "schedule my brief", "the brief didn't arrive" |
  30. Scripts (Python standard library, each with `--help`): `ticker.py`, `health_check.py`, `send_telegram.py`, `fetch_sources.py`, `carryover.py`.
  31. ## Quick start
  32. 1. Install and create a workspace:
  33. ```bash
  34. ./install.sh --workspace ~/briefing
  35. cd ~/briefing
  36. ```
  37. 2. Fill in `notes/priorities.md` (goals, open decisions, topics you track) and put your feeds in `sources.json`.
  38. 3. Ask for a brief in Claude Code, inside the workspace:
  39. ```
  40. > morning brief
  41. ```
  42. 4. When it reads right, schedule it (see `skills/brief-ticker/references/scheduling.md`) and set `TELEGRAM_BOT_TOKEN` and `TELEGRAM_CHAT_ID`.
  43. Try it offline first, with the invented coffee roaster in `examples/`:
  44. ```bash
  45. cd examples
  46. python3 ../skills/brief-ticker/scripts/ticker.py --config ticker.demo.json
  47. ```
  48. Sample summary (the full brief is in `examples/sample-morning-brief.md`):
  49. ```
  50. Morning brief 2026-03-02: 1 step(s) need attention
  51. [OK] signals: done
  52. [OK] morning-brief: done
  53. [FAIL] crm-export: capped (CRM API returned 401: token expired)
  54. [..] evening-review: pending
  55. --- Brief ---
  56. # Morning brief, Mon 2 Mar 2026
  57. System: 1 alert. Step crm-export failed (CRM token expired). Pipeline numbers below are from Friday.
  58. **#1 today:** Call Northwind Bakery before 11:00. Open 4 days; today you either call or drop the lead.
  59. ```
  60. ## How it works
  61. ```
  62. every 30 min (cron / PM2 / Task Scheduler)
  63. |
  64. v
  65. +--------------+ lock, classify each step: off / pending / due / done / missed / capped
  66. | ticker.py |--------------------------------------------------------------+
  67. +--------------+ |
  68. | 07:00 | 07:30 | 21:30 |
  69. v v v |
  70. fetch_sources.py morning-brief skill evening-review skill |
  71. (feeds -> dedupe reads priorities, notes, carryover.py -> verdicts, |
  72. -> seen-state) candidates, health_check.py tomorrow's #1 into the note |
  73. | | | |
  74. v v v |
  75. briefs/D-candidates.json briefs/D-morning.md briefs/D-evening.md |
  76. \__________ expect files prove each step ran ______/
  77. |
  78. 08:00 one summary: every step's status
  79. + brief excerpt -> Telegram or stdout
  80. ```
  81. Two rules hold it together. A step is done only when its output file exists, so "exit 0, wrote nothing" shows up as a failure. And the summary lists every configured step every day, so a skip is a visible line instead of a missing one. Longer walkthrough: [docs/how-it-works.md](docs/how-it-works.md).
  82. ## Configure
  83. | What | Where |
  84. |---|---|
  85. | Steps, times, retries, summary, health rules | `ticker.json` (from `templates/ticker.example.json`), reference in `skills/brief-ticker/references/config.md` |
  86. | Feeds and weights | `sources.json` (from `templates/sources.example.json`), feed URL patterns in `skills/signals-digest/references/sources.md` |
  87. | Goals, open decisions, hard no's | `notes/priorities.md` (from `templates/priorities.md`) |
  88. | Daily notes | `notes/daily/YYYY-MM-DD.md` (from `templates/daily-note.md`) |
  89. | Telegram | env `TELEGRAM_BOT_TOKEN`, `TELEGRAM_CHAT_ID` (see `templates/env.example`) |
  90. | Scheduler | `templates/crontab.example`, `templates/ecosystem.config.js` |
  91. ## Honest limits
  92. - The scripts are tested offline on the example data. Live feeds, Telegram delivery and the PM2 and cron templates were not run end to end by the authors on every OS.
  93. - Model steps run through `claude -p` on a schedule. That costs tokens every day, and the command and its flags may change with your Claude Code version.
  94. - No calendar or email integration is bundled. The brief uses whatever connectors your session has and says "not checked" for the rest.
  95. - Dedupe is title similarity, not embeddings. Two very differently titled articles on the same story both get through.
  96. - The scheduler uses local time and cannot wake a sleeping laptop unless you enable that in the OS. Missed steps show as `[MISS]`; they are not run later than their catch-up window.
  97. - No public feed exists for X/Twitter or most paywalled sites. Bring those in as a JSON file source.
  98. ## Credits
  99. - Brief structure ("not checked" vs "nothing found", untrusted content is data, a short reply plus a saved file): inspired by the daily-briefing skill in https://github.com/anthropics/knowledge-work-plugins (Apache-2.0).
  100. - Parallel source gathering and "scannable in 30 seconds": inspired by the chief-of-staff skill listed at https://skills.sh/hartreeworks/skill--chief-of-staff/chief-of-staff (ideas only).
  101. - Open-decisions section and flagging repeated deferrals: inspired by https://dev.to/daniel_marin_871e4c78cfc0/i-built-an-ai-chief-of-staff-it-briefs-me-every-morning-and-challenges-my-assumptions-before-i-j3m (ideas only).
  102. - Digest dedupe and Telegram delivery patterns: inspired by https://github.com/black-hak/feedmind (MIT), https://github.com/Pappet/ai_digest and https://github.com/rfrenoy/rss-reader (ideas only, no code copied).
  103. - Repo and skill layout: modelled on https://github.com/anthropics/skills and https://github.com/obra/superpowers (MIT).
  104. ## Contributing
  105. See [CONTRIBUTING.md](CONTRIBUTING.md). Agents working on the repo: [AGENTS.md](AGENTS.md).
  106. <!-- author:start -->
  107. ## Author
  108. Built and maintained by [Artem Mavrin](https://github.com/mavrin-xv). This repo is one of a set of agent skill packs for marketing and operations work. Each one installs on its own; the full list is in [agent-skill-packs](https://github.com/mavrin-xv/agent-skill-packs). If a pack saved you an afternoon, a star helps other people find it.
  109. <!-- author:end -->
  110. ## License
  111. MIT. See [LICENSE](LICENSE).

Ratings and comments

Sign in to rate this skill.

Sign in

Free to use. If it’s helped, pass it on.