Build a Custom Morning Brief with OpenClaw (News, Tasks, and Next Actions)

Most people waste their best cognitive window just getting oriented: checking calendar, scanning news, reviewing tasks, and deciding what matters. Meanwhile, your agent is idle overnight.

A custom morning brief flips that. You wake up to a structured message that already did the scanning, sorting, and first-pass thinking.

The best version is not a “daily newsletter”. It is an action plan. It shows you what matters today, what you should ignore, and what the agent can execute for you before lunch.

Internal links you may want open in a tab:

• /blog/openclaw-use-cases
• /blog/openclaw-skills-guide

What a great morning brief contains

A morning brief should be:

• Short enough to read in 2 to 4 minutes
• Specific enough to act on immediately
• Stable in structure (so you can skim)
• Personalized to your goals and constraints

A high-signal structure:

1. Today at a glance (calendar)
2. Focus plan (deep work window)
3. Top priorities (tasks)
4. Context you requested (news, metrics, messages)
5. Agent recommendations (3 concrete tasks the agent can do today)
6. One feedback question (so it improves)

The feedback loop is critical. Without it, briefs drift into noise.

Skills and integrations you may want

You can build a useful brief with only two inputs:

• calendar
• tasks

Then add the rest.

Common integrations:

• Delivery: Telegram or Discord
• Tasks: Todoist, Apple Reminders, Asana
• Calendar: Google Calendar
• Web research: built-in web search
• Optional: social research (X/Twitter trend scanning)
• Optional: email triage (Gmail)

If you are not sure how to connect integrations and permissions, read: /blog/openclaw-skills-guide

Design principles (so the brief stays valuable)

Principle 1: A strict word budget

If the brief is long, you will not read it.

Set a hard cap:

• 400 to 700 words total

If it cannot fit, the agent must drop low-value items.

Principle 2: Fixed headings every day

A stable structure turns the brief into a dashboard.

You should be able to skim by pattern:

• conflicts
• next meeting
• top task
• what the agent can do

Principle 3: “Why it matters” lines

Briefs fail when they include raw lists.

Require a one-line justification:

• for each top task
• for each news item

If the agent cannot explain relevance, it should omit the item.

Principle 4: Prefer decisions over information

Information is cheap. Decisions are expensive.

Your brief should help you decide:

• what to do first
• what to postpone
• what to delegate to the agent

Step-by-step setup

Step 1: Pick delivery channel and time

Choose one channel.

• Telegram for personal use
• Discord for team workflows

Pick a time that matches your routine. Many people choose 08:00.

Recommendation: schedule it 10 to 20 minutes before you wake up so it is waiting.

Step 2: Define your brief spec (copy-paste)

This spec prevents scope creep.

Set up a daily morning brief.

Delivery:
- Channel: Telegram
- Time: 08:00 local time

Format rules:
- Max 700 words total
- Use the same headings every day
- Prefer bullets, no long paragraphs
- Include links when referencing sources

Sections (in this order):
1) Today at a glance (calendar)
2) Focus plan (one deep work block)
3) Top tasks (max 5)
4) News (max 3)
5) Agent recommendations (3)
6) One feedback question

Personalization:
- Maintain memory of what I like and dislike in the brief
- Update the format based on my feedback
- If I do not respond, do not expand scope

Step 3: Connect tasks and calendar

A morning brief is only as useful as its “what should I do today?” section.

Minimum viable logic:

• Pull tasks due today
• Pull overdue tasks
• Pull calendar events for today
• Identify meeting clusters that block deep work

Then produce:

• Top 5 tasks ranked by impact
• One proposed deep work block
• A warning if calendar leaves no focus window

If you use Todoist, pair this with /blog/openclaw-todoist so long agent runs stay transparent.

Step 4: Define your news scope (avoid doom-scrolling)

Do not ask for “general news”. It becomes noise.

Instead:

• max 3 stories
• must match your interests
• must include “why it matters”

Example rule:

For the News section:
- Only include items related to AI tooling, dev workflows, startups, or my industry.
- If you cannot find 3 strong items, include fewer.
- Prefer sources with original reporting.
- Do not repeat the same story two days in a row unless there is a meaningful update.

Step 5: Add proactive agent recommendations

This is the most valuable section. It turns the brief into leverage.

Examples of useful recommendations:

• “I can draft the follow-up email for your 14:00 call if you paste notes.”
• “I can research 3 vendors and build a shortlist with pricing.”
• “I can turn yesterday’s meeting notes into a project update message.”
• “I can prepare an agenda and questions for your next customer call.”

Prompt:

In the Recommendations section, propose tasks that:
- you can complete end-to-end
- take less than 60 minutes each
- reduce my cognitive load today
- are grounded in my calendar and task list

For each recommendation, include:
- what you will do
- what input you need from me (if any)
- where the output will be delivered (message or file path)
- what "done" looks like

Step 6: Add an execution shortcut

Make recommendations actionable in one message.

Rule:

• If you reply “Do #2”, the agent executes recommendation 2.

Prompt:

Add a control rule:
- If I respond with "Do #N", immediately execute recommendation N.
- Confirm start, then report completion with the output.
- If you need extra input, ask one question.

Step 7: Install the feedback loop

Every brief should end with one question.

Examples:

• “Was this too long or too short?”
• “Do you want more tasks or more context tomorrow?”
• “Which section should I remove?”

Then store preferences:

• preferred brief length
• preferred news topics
• excluded topics
• preferred task prioritization style

Recommended output format

Keep it consistent. Example skeleton:

• Today at a glance

  • 09:30 Product sync (Zoom)
  • 14:00 Customer call (Office)

• Focus plan

  • 10:30 to 12:00 deep work: ship onboarding copy integration

• Top tasks

  • Finish proposal draft (unblocks pricing decision)
  • Review PR #123 (reduces deployment risk)
  • Prep questions for 14:00 call (improves call outcome)

• News

  • Link: item (why it matters)

• Agent recommendations

  • #1 Draft follow-up email for 14:00 call (needs notes, outputs in Telegram)
  • #2 Research 3 competitor onboarding flows (outputs in markdown file)
  • #3 Turn yesterday’s notes into a project update (outputs to Discord)

• One question

  • “Should I reduce news to 2 items tomorrow?”

Personalization that actually matters

Most personalization requests are shallow (add quotes, add emojis). Focus on what changes decisions.

Personalization lever 1: weekly goals

Give the agent a weekly goals list. Then the brief ranks tasks against those goals.

Example:

My goals this week:
1) Ship onboarding improvements
2) Publish 2 blog posts
3) Close 3 sales calls

In the brief, rank tasks based on these goals.

Personalization lever 2: “do not include” list

A short exclusion list can double signal.

Example:

• Do not include general politics
• Do not include crypto price chatter
• Do not include product launches unless they impact my stack

Personalization lever 3: your work style

If you hate context switching, add a “batching” rule:

• group tasks by theme
• propose one theme per day

Common upgrades (when the base brief works)

Upgrade 1: Add weather and commute context

If your day depends on travel:

• weather summary
• commute time buffer

Keep it to one line. If it does not change logistics, omit it.

Upgrade 2: Add “what changed since yesterday”

This prevents repetition. Great for:

• KPIs
• git commits
• support tickets

Upgrade 3: Add a “stuck items” section

If a task is overdue for 7 days, the brief should ask:

• “Do you want to drop this, delegate it, or schedule it?”

Upgrade 4: Add a “one big bet” suggestion

Once a week, the agent suggests one high-upside action:

• partnership idea
• outreach campaign
• product experiment

Keep it separate from daily execution.

Troubleshooting

The brief is too long

Fixes:

• enforce a hard word limit
• reduce news from 3 to 2
• reduce tasks from 5 to 3
• remove one optional section

The brief is irrelevant

Fixes:

• narrow news scope
• provide weekly goals
• enforce “why it matters” lines

The brief repeats the same stories

Fixes:

• maintain a small “seen items” list for the last 7 days
• add a rule: do not repeat unless materially updated

Recommendations are unrealistic

Fixes:

• constrain to tasks under 60 minutes
• require “inputs needed” and “definition of done”
• if inputs are missing, the agent must ask one question

FAQ

Q: Daily or weekdays only? A: Weekdays only reduces noise. If you work weekends, keep it daily or switch to a lighter weekend format.

Q: Do I need social research skills? A: Not required. Start with calendar + tasks + web news. Add X/Twitter scanning only if it consistently produces actionable insights.

Q: How do I prevent it from turning into a news feed? A: Keep news to 2 or 3 items and require “why it matters”. If it cannot justify relevance, it must omit the story.

Q: Can I make the brief trigger automations automatically? A: Yes, but start conservative. A good pattern is “manual trigger”: you reply “Do #2”. Automatic triggers can be added later for high-confidence tasks.

Q: How do I personalize it quickly? A: Give it weekly goals and an exclusion list. Personalization is mostly constraint setting.

Q: How does this relate to other OpenClaw use cases? A: Morning briefs are a glue workflow. They surface the top items from task visibility, project tracking, household assistants, and multi-agent teams. Explore more at /blog/openclaw-use-cases.

Examples by persona (copy and adapt)

If you want faster results, tailor the brief to your role.

Solo founder

• News: funding, product launches, AI tooling
• Tasks: one product task, one distribution task, one sales task
• Recommendations: “draft outreach”, “analyze funnel”, “write changelog post”

Add a rule:

• “Always include one action that increases distribution today.”

Engineer with meetings

• Calendar: highlight meeting clusters
• Focus plan: propose a 90 minute coding window
• Recommendations: “prepare meeting agenda”, “write PR description”, “summarize incident report”

Add a rule:

• “If calendar has no focus window, propose one meeting to decline or move.”

Creator or marketer

• News: competitor campaigns and channel trends
• Tasks: one content asset, one repurposing action, one distribution action
• Recommendations: “draft thread”, “outline video”, “collect sources”

Add a rule:

• “Include one finished draft, not just ideas, at least twice per week.”

A simple 3-day iteration plan

Day 1:

• Ship the brief with calendar + tasks only.
• Reply with one constraint you want.

Day 2:

• Add recommendations and the “Do #N” execution shortcut.
• Remove any section you did not read.

Day 3:

• Add a narrow news scope (max 2 or 3 items) and enforce “why it matters”.
• Lock the word limit.

After day 3, do not expand scope. Improve relevance and consistency instead.

Next step

Once your brief is stable, connect it to execution:

• when the brief proposes a task, the agent can create a Todoist “agent run” task and stream progress (see /blog/openclaw-todoist)
• include project status pulled from your event-driven tracker (see /blog/openclaw-project-tracking)

Want a morning brief that is actually actionable, not just a summary? Deploy OpenClaw with the form above, paste the spec prompt, and iterate for three days. The feedback loop is where the value compounds.