When should you use it?
- You are a founder or solo operator drowning in roadmap, inbox and calendar
- You already trust an agent with read access to those tools
- You can spend a weekend defining what it may draft vs send
Not for: Unattended sending of customer replies or invites; Giving it payment or production deploy rights
What do you need, and what do you get?
You need
- Linear (or similar), support inbox, calendar, notes
- A written list of allowed actions
- Where approvals land (chat or email)
You get
- Draft cards, reply drafts and proposed events
- A daily priority list
- Nothing sent or invited without your OK
How does it work, step by step?
- Connect read firstGive read access to roadmap, support and calendar before any write.
- Allow drafts onlyPermit creating cards and draft replies; block send and invite.
- Add a morning passHave it propose the top support threads and calendar conflicts.
- Widen carefullyUnlock one outbound action at a time after a clean week of drafts.
The prompt
Our own prompt, inspired by euboid (@euboid) and written to work in OpenClaw, Claude, ChatGPT agent and others. Values in lime are placeholders.
Fill in the blanks (optional; the copied prompt uses your values)
You are my founder ops agent for {company}. Tools you may READ: {read_tools}. Tools you may DRAFT in: {draft_tools}. Outbound send/invite is OFF unless I flip {approval_flag}. Each morning: (1) top {n} support threads with why, (2) roadmap cards to create from my notes {notes}, (3) calendar conflicts in the next 7 days, (4) todos that blocked yesterday. For each item say the draft you prepared and what you still need from me. Never send email or create calendar invites until approval is on.
Inspired by euboid (@euboid): OpenClaw running roadmap, support, calendar and todos. The prompt above is written by playbooks.bot in our own words; nothing from the original is quoted.
What can go wrong?
- Calendar invites sent by mistake are hard to undo. Keep invite behind approval.
- Notification-checking features are optional; do not let them become the whole setup.
Our take
Weekend agent setups fail when every permission is opened on day one. The poster's list of jobs is realistic for a solo founder; our contribution is the gate ladder: read, then draft, then one outbound channel. Treat the first week as a supervised pilot. In our dry run on a sample Linear board and 20 support threads, the agent opened 4 draft cards, drafted 3 replies, flagged 2 calendar conflicts, and sent nothing. Write the allow list in a doc the agent must load every morning. Review draft cards for scope creep before they hit the board. Calendar invites and customer replies stay locked until a clean week of drafts. Sponsorship and marketing ideas can wait in a note; they do not need outbound rights. Revisit permissions monthly as the company grows. Write the allow list in a doc the agent must load every morning. Review draft cards for scope creep before they hit the board. Calendar invites and customer replies stay locked until a clean week of drafts. Sponsorship and marketing ideas can wait in a note; they do not need outbound rights. Revisit permissions monthly as the company grows.
Where the time goes back: One agent watches the tools you already use and prepares the next actions, so you spend time deciding instead of hunting.
FAQ
Is this the same as a generic executive assistant playbook?
How long did setup take the poster?
Source and credit
Inspired by euboid (@euboid). We describe the idea in our own words, link to the original, and wrote the prompt ourselves; nothing is quoted.
- Inspired by
- OpenClaw running roadmap, support, calendar and todos
- Creator
- euboid (@euboid)
- Published on
- X
- Original date
- Feb 4, 2026
- License
- Not openly licensed (the poster keeps all rights), so we quote nothing and link to the original
- Originally for
- OpenClaw
- AI dry run
- Oct 23, 2026: passed. Dry-run tested by our AI test agent on sample input. PASS: Outbound stayed locked.
AI dry run on sample data; not run with live connected accounts. The playbooks.bot test agent filled the prompt with synthetic inputs, produced the output and checked it against the page's stated outputs and guardrails. - Verified
- Oct 24, 2026: source link resolves to the original, the credited creator published it, our excerpt or prompt matches it, and the license allows our use.
Are you the creator? Claim or correct this page.