Get Your Own /today

A daily AI-powered intelligence briefing system built on Claude Code. Ten specialized agents gather news, research, social trends, legal developments, and your calendar — then assemble everything into a personalized briefing with social posts and an interactive dashboard.

What You Get

Every morning, one command (/today) runs a 10-agent pipeline:

1. Research Agent — Pulls academic papers from arXiv matching your interests
2. Calendar Agent — Fetches your schedule (supports Google, Outlook, Apple, or none)
3. Intel Agent — Scans 20+ news sources, deduplicates against yesterday
4. Social Agent — Monitors influencers across 6 platforms, scores trends
5. Legal Agent — Searches court filings and rulings in your domain
6. Briefing Agent — Assembles everything into a structured daily briefing
7. Social Posts Agent — Generates ready-to-publish posts from today's trends
8. Self-Update Agent — Learns from each run, tunes sources and scoring
9. Dashboard Agent — Regenerates an interactive HTML dashboard
10. Publish Agent — Pushes to GitHub/Vercel (optional)

The system improves itself daily — bad sources get paused, new ones get discovered, scoring weights adapt.

Quick Start

Prerequisites

• Claude Code installed and working
• Python 3.10+ with pip
• A topic domain you care about (the setup wizard will ask)

Install

# 1. Copy this folder as "daily-briefing" into your working directory
cp -r DISTRO /path/to/your/project/daily-briefing

# 2. Copy the Claude Code command so /today works
mkdir -p /path/to/your/project/.claude/commands
cp daily-briefing/.claude/commands/today.md /path/to/your/project/.claude/commands/today.md

# 3. Copy the Claude Code permissions
cp daily-briefing/.claude/settings.local.json /path/to/your/project/.claude/settings.local.json

# 4. Install Python dependencies
pip3 install arxiv pyyaml requests

# 5. Set up your environment
cp daily-briefing/.env.example daily-briefing/.env
# Edit .env with your API keys

# 6. Run it
cd /path/to/your/project
claude
# Then type: /today

On first run, the system detects empty configs and walks you through setup:

• Your domain — What industry/topic do you want intelligence on?
• Your calendar — Which provider? (Google Calendar works out of the box with Claude Code's built-in MCP. Outlook needs Azure credentials. Apple Calendar uses an ICS feed. Or skip it.)
• Your voice — How do you write on social media?
• Your mission areas — What products/projects should intel map to?

It then auto-generates starter configs (news sources, influencers, research keywords, legal queries) tailored to your domain and runs the first briefing.

Environment Variables

Variable	Required?	Used By
ANTHROPIC_API_KEY	Yes (Claude Code needs it)	Runtime
FIRECRAWL_API_KEY	Recommended	Intel, Social, Legal agents (web scraping)
OPENROUTER_API_KEY	Optional	Social post image generation
SLACK_WEBHOOK_URL	Optional	Post intel to Slack
VAULT_PASSPHRASE	Optional	Dashboard password protection

Calendar credentials are configured during the first-run setup wizard.

How It Works

config/*.yaml → [5 gather agents] → output/.staging/*.json → [assembly agents] → output/briefings/ + output/dashboard/
                                                                      ↓
                                                             config/*.yaml (self-update writes back)

Each agent reads a config, fetches data, writes a staging JSON with a standard contract, and reports its status. The briefing agent combines everything. The self-update agent makes the whole system smarter over time.

Mission Board

While running, you see a live status display:

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
 DAILY BRIEFING — 2026-04-02 — IN PROGRESS
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

 PHASE 1: GATHER INTELLIGENCE
 ├─ 🟢 Research         ✓  14 papers pulled
 ├─ 🟢 Calendar         ✓  5 events loaded
 ├─ ⏳ Intel            …  Fetching sources (8/13)
 ├─ ⬜ Social           ·  Waiting
 └─ ⬜ Legal            ·  Waiting

 PHASE 2: ASSEMBLE
 ├─ ⬜ Briefing         ·  Waiting
 └─ ⬜ Social Posts     ·  Waiting

 PHASE 3: LEARN & PUBLISH
 ├─ ⬜ Self-Update      ·  Waiting
 ├─ ⬜ Dashboard        ·  Waiting
 └─ ⬜ Publish          ·  Waiting
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

Recommended Setup: Obsidian + Claude Code Sidebar

This system works anywhere Claude Code runs. But the optimal setup is inside an Obsidian vault with the Claude Code sidebar extension.

Why Obsidian:

• Your daily briefings are markdown files — Obsidian renders them beautifully with wikilinks, tags, and backlinks
• Research papers, social posts, and briefings all cross-reference each other via [[wikilinks]]
• The briefing becomes part of your knowledge graph, not a throwaway document
• Inspiration images embed inline with ![[image.png]]
• Tasks, wins, and ops logs are just markdown files you can search, link, and organize

Why the Claude Code sidebar:

• Run /today without leaving your vault — the sidebar opens Claude Code right inside Obsidian
• Watch the mission board update in real time while you read yesterday's briefing
• Click through from the briefing to social posts, dashboard HTML, or config files without switching apps
• Edit your voice.md or mission.yaml in one pane, run /today in the other

Setup:

1. Install Obsidian and create a vault (or use an existing one)
2. Install the Claude Code sidebar extension from Obsidian's community plugins
3. Drop daily-briefing/ into your vault folder
4. Copy .claude/commands/today.md to your vault's .claude/commands/
5. Open the Claude Code sidebar and type /today

Your briefings accumulate as daily notes. Your configs evolve over time. Your research papers build a library. It's not just a briefing system — it becomes your second brain's morning routine.

Not required. Just optimal.

Customization

Change your domain

Edit these files to shift to any industry:

File	What to change
config/mission.yaml	Your products/projects and mission areas
config/intel_sources.yaml	News sources for your industry
config/social_sources.yaml	Influencers and companies to track
config/research_keywords.yaml	arXiv search terms
config/legal_queries.yaml	Court search queries
config/voice.md	Your writing style for social posts

Add/remove agents

Each agent is a standalone markdown file in commands/agents/. The orchestrator (commands/today-v2.md) reads and executes them in order. Add a new .md file, add it to the orchestrator's phase list, done.

Architecture

Agent	Config File	Output
research	config/research_keywords.yaml	research.json
calendar	config/calendar_config.yaml	calendar.json
intel	config/intel_sources.yaml	intel.json
social	config/social_sources.yaml	social.json
legal	config/legal_queries.yaml	legal.json
briefing	config/briefing_prefs.yaml	briefing-meta.json
social-posts	config/post_performance.yaml	posts-meta.json
self-update	All configs	updates.json
dashboard	Templates	Dashboard HTML
publish	scripts/sync_dashboard.sh	ops.json

Cost

Each /today run costs pennies in API credits. A full month of daily runs has been measured at under $0.50 total using Claude Haiku. Cost varies by model — Opus will cost more, but even heavy usage stays well under $5/month.

Credits

Built by Phil Komarny using Claude Code. Open-sourced so anyone can have their own AI-powered daily briefing.