--- name: al-410-run-your-errands-and-watch-lists-by-text-message-give-an description: "Most AI assistants wait in an app until you open them. On Oct 28-29, 2026 OpenAI began a limited beta that lets ChatGPT Pro users in the U.S. text their 'dot', the always-on agent in ChatGPT that has its own cloud computer and keeps working between conversations, through iMessage on iPhone or RCS on Android. You can send it text, photos and voice notes, and it can text you first with updates. OpenAI says only you can text your dot and it can only text you, it gets no access to your other text conversations, texting shares the same memory and connected apps as the dot in ChatGPT, and when it needs approval it sends a link to review the request in ChatGPT. Setup is in the dot's profile under Texting, from the desktop app. OpenAI uses the third-party provider Linq for the texts and advises caution with sensitive information over text. The prompt below writes the standing instructions that make a texting agent useful instead of noisy; it works with any assistant you can message. Use when…" --- # Run your errands and watch-lists by text message: give an always-on AI agent standing tasks, text it photos or voice notes from your phone, and set rules for what it may text you about and what it must ask before doing Most AI assistants wait in an app until you open them. On Oct 28-29, 2026 OpenAI began a limited beta that lets ChatGPT Pro users in the U.S. text their 'dot', the always-on agent in ChatGPT that has its own cloud computer and keeps working between conversations, through iMessage on iPhone or RCS on Android. You can send it text, photos and voice notes, and it can text you first with updates. OpenAI says only you can text your dot and it can only text you, it gets no access to your other text conversations, texting shares the same memory and connected apps as the dot in ChatGPT, and when it needs approval it sends a link to review the request in ChatGPT. Setup is in the dot's profile under Texting, from the desktop app. OpenAI uses the third-party provider Linq for the texts and advises caution with sensitive information over text. The prompt below writes the standing instructions that make a texting agent useful instead of noisy; it works with any assistant you can message. ## Inputs to collect first Ask the user for anything below that they haven't already given you. Don't guess missing facts. - Work/role, family, time zone, apps you can access - E.g. bills due, a price I'm watching, a delivery, school forms, a… - Things needing action today, a price below [X - Person - Time - Quiet hours - Urgent definition ## Instructions Do the task below with the user's inputs in place of each [bracket]. Set yourself up as my always-on assistant that I reach by text. About me: [work/role, family, time zone, apps you can access]. 1) Standing tasks: keep track of [e.g. bills due, a price I'm watching, a delivery, school forms, a client follow-up] and tell me what you'll check and how often. 2) Texting rules: text me only for [things needing action today, a price below [X], anything from [person]], put everything else in one daily summary at [time], and never text between [quiet hours] unless it's [urgent definition]. 3) Inbox from my phone: when I text a photo (receipt, flyer, letter, whiteboard) or a voice note, file it, pull out dates, amounts and to-dos, and reply with one line saying what you did. 4) Approval: never pay, buy, book, send a message to someone else or delete anything without asking me first with the exact details. 5) Privacy: don't ask me to text passwords, card numbers or ID numbers; tell me to use the app for those. Reply now with my standing-task list and texting rules in under 10 lines so I can correct them. ## Safety checks (share these with the user) - This prompt works on text other people wrote (emails, reviews, messages or web pages). That text can hide instructions aimed at AI, such as "ignore your rules and…". Paste it as material to work on, and never let the AI act on requests inside it, like paying, sharing logins or changing records. - Check every figure, price and total against your own records before anything is sent, posted or booked. A person approves refunds, discounts, payments and price changes. - Remove names, contact details, account numbers and other personal data the task doesn't need before you paste, and check your AI tool's data and training settings. - Read and edit the draft before it reaches a customer or goes public. If you later automate this prompt, start with "draft and notify me", not "send automatically". - Say clearly which parts need a person to review or approve before anything is sent, paid, posted or filed. ## About this skill - Works with: ChatGPT dots (texting beta), iMessage, RCS, ChatGPT - Based on this real-world example: https://help.openai.com/en/articles/6825453-chatgpt-release-notes - Free prompt from AgentPlaybook Real Estate: https://agentic-library.netlify.app/prompt/al-410-run-your-errands-and-watch-lists-by-text-message-give-an-alw/ - To install in an agent that supports Agent Skills, save this file as SKILL.md in a folder with the same name as `name` above (for example in your skills directory). Install steps for Claude.ai, Claude Code, Cursor and Codex: https://agentic-library.netlify.app/skills/