Weekly Summary Workflow Prompt You Run Yourself

Start here

This is a plain-text prompt for an ordinary AI chat that accepts text. You do not need a terminal, API, account setup or any add-on.

What to paste. First paste this original prompt, then paste your own short list: the summary's goal and readers, the places the week's material lives (notes app, shared doc, chat thread, spreadsheet, folder), the exact start and end dates, the section order you want, any recurring callouts, length and tone limits, and last week's summary if you have one. Write MISSING for anything you don't have.

Where to paste. Into any general chat assistant that can read pasted text. Keep it in one conversation so the assistant can refer back.

How to check the output. The reply should contain a source table where every source has a concrete manual retrieval step, a numbered retrieval checklist, one copyable weekly prompt block, file-name guidance, 6-10 pass/fail review checks, missed-week steps and a short boundaries list. If a source has no manual retrieval path, it should be flagged and stopped, not guessed. The weekly prompt inside the reply must require a source label on every claim, mark single-source items as SINGLE-SOURCE and unclear ones as UNCLEAR.

What this is for

Weekly summaries fail when the person writing them expects software to collect, schedule or send the material. This prompt removes that assumption. It only produces two things: a plan you can carry out by hand, and a reusable prompt you paste each week with the week's material underneath.

Note the word fictional in the prompt's example. It only shows the expected shape; the numbers and outcomes in it are not test results. Source reviewed; runtime not tested.

What you get back

Permissions and limits

• The plan is executed by you: every export, copy and send is a manual action.
• Nothing here schedules, watches files, sends messages or remembers previous weeks.
• If the goal, sources or time window are missing, the assistant should ask before writing the prompt. Style-only gaps are labelled as assumptions.
• Do not paste confidential material into the chat unless the provider's terms and your own confidentiality limits allow it. Redact names, figures or client details first.
• Treat the drafted summary as a draft: you check the quotes, numbers and labels before sharing.

FAQ

Do I need an app or automation tool? No. The whole point is that the workflow runs through your own hands. The prompt explicitly forbids claiming that any tool can watch, schedule or send for you.

Can it recover a week I skipped? It gives numbered steps: extend the window, label the gap, and keep two weeks from being merged into one. You still copy the material yourself.

Complete reusable prompt

You are helping me design a recurring weekly-summary workflow that I will run myself. Do not assume any tool can schedule itself, watch my files, send messages, or remember last week. I will supply the inputs; you only produce a plan and a reusable prompt I can paste into a chat later.

INPUTS I WILL PROVIDE (paste what you have; write MISSING for the rest)

1. Goal: what the weekly summary is for, who reads it, and what decision or action it supports.
2. Sources: the places the week's material lives (for example a notes app, a folder of meeting notes, a spreadsheet, an inbox label, a chat thread, a shared doc). For each source, say how I get its content out by hand or by export.
3. Time window: the exact start and end, plus the day I will assemble it.
4. Sections I want in the summary and their order.
5. Known recurring items (callouts that appear most weeks).
6. Constraints: length, tone, anything that must never appear, confidentiality limits.
7. Last week's summary, if one exists, so you can match its shape.

WHAT TO DO Step 1 - Sources check. Build a table: Source | How I retrieve it | Format it arrives in | Anything I must do first (e.g. export, rename, copy). If any source has no manual retrieval path, flag it and propose a substitute, then stop that source rather than guessing. Step 2 - Assembly order. Give me a numbered retrieval checklist in the order I should copy or export material, with a one-line purpose for each step. Keep it to the minimum that satisfies the goal. Step 3 - Reusable summary prompt. Write me one self-contained prompt I can paste into a chat each week, with the pasted material underneath. That prompt must: state the section order; require every claim to carry the source label I supplied; mark any item that appears in only one source as SINGLE-SOURCE; mark anything unclear as UNCLEAR rather than filling it in; keep to my length and tone limits; and end with a short list of open questions for me. Step 4 - Draft pack. Tell me what to keep after each run (the pasted inputs, the drafted summary, my edits) and how to name files so a later week can be compared without re-reading everything. Step 5 - Weekly review checklist. Give me 6-10 pass/fail checks I run on the drafted summary before I send or share it, tied to my stated goal and sources, not generic advice. Step 6 - Missed week. Describe what happens if I skip a week: how to extend the window, how to label the gap, and how to avoid presenting two weeks as one. Step 7 - Boundaries. State plainly what this workflow does NOT do: it does not schedule, monitor, send, or remember anything by itself; all retrieval and any sending is my manual action.

OUTPUT FORMAT

• 'Setup' with the sources table and the retrieval checklist.
• 'Reusable weekly prompt' as one clearly delimited block of plain text I can copy.
• 'Keep and name' short bullet list.
• 'Review checklist' as checkbox items.
• 'If I miss a week' short numbered steps.
• 'Boundaries' three to five bullets.

MISSING-INPUT RULE If goal, sources, or time window is missing, ask me for them before producing the prompt. If only style details are missing, proceed and label your assumption clearly at the top.

FICTIONAL EXAMPLE (for shape only) Input: Goal - a short Friday recap for a three-person study group. Sources - our shared notes doc and a group chat. Window - Monday to Friday. Sections - done, blocked, next week, questions. Expected shape: the retrieval checklist says 'open notes doc, copy the week's headings'; the reusable prompt returns four sections, each line labelled with its source, blocked items flagged UNCLEAR where the chat and notes disagree, and two open questions at the end.

CHECKS BEFORE YOU HAND ME THE RESULT

• Does every source in my list have a concrete manual retrieval step?
• Does the reusable prompt require source labels on every claim?
• Does it flag one-source and unclear items instead of smoothing them over?
• Is the whole plan runnable by me alone, with no assumed automatic scheduling or sending?
• If last week's summary was supplied, does the shape match it?

BOUNDARIES Keep to my supplied material. Do not invent sources, numbers, or recurring items. Do not claim any tool can watch, schedule, or send for me. Preparing the plan and the draft prompt is not the same as running or sending the summary; that part stays manual.

References and reuse

• ChatGPT release notes · Reviewed 2026-10-04

Original TokRepo prompt · CC BY 4.0. Reference documents retain their own rights.