← Back to library

Make every agent save finished work where you can find it

Michael Fenech (@Michael_Fenech_) posted a setup prompt for Grok Bot teams that keeps finished files from scattering: agree a home for each output type, use clear names, update the original instead of making copies, and hand back a working link. We adapt it for any team of agents writing to shared drives.

  • OutcomeAgree one home and one naming pattern for the documents, spreadsheets and reports your agents produce, add that rule to every agent, and get a link plus a one-line status after each save.
  • SetupMinutes
  • LevelStarter
  • CreditMichael Fenech (@Michael_Fenech_)
View original
Works with
Grok BotClaude CoworkChatGPT agentMicrosoft Copilot

Published Oct 27, 2026Updated Oct 27, 2026AI dry run on sample input: passed Oct 27, 2026Source verified Oct 27, 2026

When should you use it?

  • Agents produce files and you keep searching for the latest version
  • Different bots save to different places or name things inconsistently
  • You share agent output with clients or colleagues and need links that work

Not for: Reorganizing or moving your existing archive; this rule covers new output; Storing passwords, contracts with signatures or regulated records without your normal controls; Single one-off chats where nothing gets saved

What do you need, and what do you get?

You need

  • The drives or folders you already use (for example OneDrive, Google Drive or SharePoint)
  • The types of output your agents produce: documents, sheets, reports
  • Permission for the lead agent to update other agents' instructions, or a list of where to paste the rule

You get

  • A short written rule: one location per output type and a naming pattern
  • Updated instructions on each agent and routine, with a list of any it could not update
  • A handover format: file name, direct link, and one sentence on what is ready or blocked
  • Two sample files saved by the rule, with links for you to open

How does it work, step by step?

  1. Pick the homesChoose one folder per output type, starting with places you already use, and confirm the agent can write there.
  2. Set the naming patternProject plus purpose, with a date for recurring reports. Revisions update the agreed file rather than adding a copy.
  3. Define the handoverAfter each save the agent returns the name, a direct link and a one-line status, and checks the file exists.
  4. Roll it outAdd the rule to every agent and routine; ask for a list of anything it could not change.
  5. Test with samplesHave it create one document and one spreadsheet, then open both links yourself.

The prompt

The original prompt by Michael Fenech (@Michael_Fenech_), word for word, with a link to where it was published. Below it is our adapted version with fill-in blanks, if you want it. Works in Grok Bot, Claude Cowork, ChatGPT agent and others. Values in lime are placeholders.

Original prompt by Michael Fenech (@Michael_Fenech_)

Word for wordby Michael Fenech (@Michael_Fenech_)
Set up one filing rule for the work my Bot team creates.

1 - Agree where things belong.

Ask where I want documents, spreadsheets and reports saved. Start with the places I already use and check you can save there. Keep it simple: one agreed home for each type of output.

2 - Make the right version easy to find.

Use clear names that identify the project and purpose. Add a date for recurring reports. When revising an existing item, update the agreed original where possible instead of creating another competing copy. Preserve its history where supported.

3 - Return a useful handover.

After saving, give me the item’s name, direct link and a sentence explaining what’s ready or what needs my attention. Check the saved item exists. If saving fails, tell me what remains unsaved.

4 - Apply the rule across the team.

Add the agreed filing instructions to the relevant Bots and routines you can update. Show me where you saved the rule and flag anything you couldn’t apply. Leave existing files where they are.

5 - Test it.

Create one sample document and one sample spreadsheet. Save them in the agreed places and return their links so I can check the result.

Exactly as published by Michael Fenech (@Michael_Fenech_) in X post: one filing rule so finished Grok Bot work lands where you expect.

Our adapted version (optional)

Fill in the blanks (optional; the copied prompt uses your values)

Adapted prompt2 variables
Create a single saving rule for files my agents produce.

- Ask me which location to use for each output type ({output_types}), starting with {current_drives}. Confirm you can write to each.
- Naming: project, then purpose; add the date only for recurring reports. When revising, overwrite the agreed file if version history is on, rather than saving a second copy.
- After every save, reply with the file name, a direct link and one sentence on what is done or needs me. Confirm the file is really there; if the save failed, say what is missing.
- Add this rule to each agent and scheduled job you can edit and list any you could not. Do not move existing files.
- Test by saving one sample document and one sample sheet and sending me both links.

What can go wrong?

  • Updating the original only works where the drive keeps version history. Check that first.
  • If an agent reports a save without a working link, treat the save as failed.
  • Leave existing files alone; a cleanup pass is a separate, reviewed job.
  • Shared folders inherit sharing settings. Do not point client-facing output at a folder anyone can open.

Our take

Agents are getting good at producing work, and that creates a new chore: finding it. Without a rule, each bot picks its own folder and naming style, revisions arrive as fresh copies, and the person who asked for a report ends up searching three drives. Fenech's prompt is unglamorous in a good way. It agrees one home per output type, a naming pattern with dates only where they help, revisions that update the original, and a handover with a direct link and a one-line status. The insistence on checking that the saved file exists is the detail most setups miss; agents sometimes report success when an upload failed. Two cautions. Updating the original assumes your drive keeps version history, so confirm that before relying on it, and sharing settings follow the folder, so keep client work out of open folders. In our dry run with a fictional two-person agency using OneDrive (a proposals bot, a weekly metrics routine and a research bot), the agent proposed three folders and a pattern like 'ClientName - Proposal - v1', saved a sample document and spreadsheet with links, and listed one routine it could not update because its instructions were locked.

Where the time goes back: You stop hunting for agent output and stop comparing near-identical copies, because every file lands in a known place with a name you recognize and a link in the handover.

FAQ

Why insist on a link after every save?
A link proves the file exists where it should. A message saying the report is done, with nothing to click, is how files get lost.
Does this work with a single agent?
Yes. It matters more as you add agents, but one assistant saving to random folders causes the same problem.

Source and credit

The original prompt on this page is by Michael Fenech (@Michael_Fenech_), shown word for word with a link to where it was published. The summary, notes and adapted version are ours.

Original
X post: one filing rule so finished Grok Bot work lands where you expect
Creator
Michael Fenech (@Michael_Fenech_)
Published on
X
Original date
Oct 26, 2026
License
Not openly licensed; the original prompt is shown word for word with credit and a link to the original
Originally for
Grok Bot
AI dry run
Oct 27, 2026: passed. Dry-run tested by our AI test agent on sample input. One home per type, link-backed handover, locked routine reported rather than skipped silently.
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 27, 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.

More in Operations & Admin

All 203 Operations & Admin playbooks →

Related playbooks

Operations & AdminPrompt✓ AI dry run

Decide whether an overloaded agent should be split in two

Have your lead agent review two weeks of instructions, routines, task history and corrections for every agent, find where work actually breaks down, separate overload from fixable causes like vague instructions or missing access, give each agent a keep, simplify or split verdict, and propose one measured trial before anything changes.

Credit: Michael Fenech (@Michael_Fenech_)

Under an hour
Operations & AdminPrompt✓ AI dry run

Give email-writing agents six standards and a learning rule

Have your lead agent apply one shared drafting process to every agent that writes or sends email (purpose, context, tone, accuracy, clarity and a pre-send check), learn from the edits you make, and propose instruction changes that you approve before they take effect, without touching existing sending permissions.

Credit: Michael Fenech (@Michael_Fenech_)

Under an hour
Operations & AdminPrompt✓ AI dry run

Give your agents office hours instead of constant pings

Have your lead agent list every agent and routine that messages you, sort each message type into urgent, decision or FYI, batch decisions into two daily briefings with a recommendation each, collect routine updates into a single evening summary, and set clear rules for the rare cases that may interrupt you.

Credit: Michael Fenech (@Michael_Fenech_)

Under an hour
Operations & AdminPrompt✓ AI dry run

Run each project from one page your agents keep current

Have your lead agent give every active project its own living page (goal, status, owners, deadlines, decisions, blockers and a change log), keep a master list of all projects, require every agent to read the page before working and update it afterward, and run a weekday check that only pings you about blocked, late or decision items.

Credit: Michael Fenech (@Michael_Fenech_)

Under an hour
Operations & AdminPrompt✓ AI dry run

Find routines your agents run twice and merge them

Have one agent collect every scheduled routine across your agent team, compare the saved instructions of any that look alike, flag pairs that read identical sources or run too close together, and hand you a short keep, merge or cut table with the expected saving, without changing anything until you say yes.

Credit: Michael Fenech (@Michael_Fenech_)

Minutes