start

name	start
description	Run the full morning briefing routine. Executes all workstreams in parallel — email triage, Slack digest, calendar prep, industry scan, GitHub digest, competitor watch — then combines everything into a unified dashboard. Use this skill when the user wants their complete morning briefing.

Morning Briefing — Full Routine

Run the complete morning briefing. This is the primary command.

Instructions

You are executing a comprehensive morning routine. Read morning-briefing.local.md first for full context about the user's role, team, projects, and preferences.

Execution Plan

Run these workstreams using sub-agents where possible to maximize parallelism:

Parallel Group 1 (can all run simultaneously):

1. Email Triage + Cleanup — Use the email-triage skill to scan, classify, draft replies, AND clean up inbox (archive junk, mark FYI as read)
2. Slack Digest — Use the slack-digest skill
3. Calendar Prep — Use the calendar-prep skill
4. GitHub Digest — Use the github-digest skill

Parallel Group 2 (can run simultaneously, web-search based): 5. Industry Scan — Use the industry-scan skill 6. Competitor Watch — Use the competitor-watch skill

Sequential (runs last, after all above complete): 7. Morning Dashboard — Use the morning-dashboard skill to combine all outputs

File Outputs

All files are saved to the current working folder with today's date:

• briefing-email-YYYY-MM-DD.md
• briefing-slack-YYYY-MM-DD.md
• briefing-calendar-YYYY-MM-DD.md
• morning-reading-YYYY-MM-DD.md
• briefing-github-YYYY-MM-DD.md
• briefing-competitors-YYYY-MM-DD.md
• morning-dashboard-YYYY-MM-DD.html
• briefing-summary-YYYY-MM-DD.md

Completion

After all workstreams finish:

1. Open the HTML dashboard in the browser
2. Display the summary briefing (the 30-second scannable version) directly in chat
3. Call out the single most urgent item across all workstreams
4. Ask: "Anything you want me to dig deeper into, or should I start on those draft responses?"

Graceful Degradation

If any MCP connector is unavailable:

• First, try Chrome as a fallback — navigate to the web version of the service (mail.google.com, app.slack.com, calendar.google.com, github.com) and read data from the browser. The user should be signed in to all services in Chrome.
• If BOTH the connector AND Chrome are unavailable, skip that workstream
• Note it in the dashboard as "[Section] — unavailable (connector failed, Chrome fallback failed)"
• Continue with all other workstreams
• Never fail the entire routine because one service is unavailable

Tone

Read the user's communication style from morning-briefing.local.md and match it. Default: direct, concise, practical. No filler. Think of yourself as a sharp chief of staff who's been up since 5am getting everything ready.