For about a year my standup routine was a small daily humiliation. I'd open Notion at 9:55am, four minutes before the call, and stare at a wall of half-sentences and emoji shortcuts I'd typed the night before. 'fixed the flaky thing — auth?? race cond. added sleep (gross). pinged raj re: migration, no reply.' Then I'd panic and wing it. My updates were either too vague — 'worked on the dashboard stuff' — or so painfully detailed that I'd watch three people's eyes glaze in real time on the Zoom grid while I narrated the entire saga of a single test failure.
The annoying part is that I always had the information. I'm a compulsive note-taker; my problem has never been knowing what I did. The problem was translation: turning a private shorthand that only I understood into something ten other humans could absorb in the fifteen seconds of attention a standup actually gets before someone's mind wanders to lunch.
Why raw notes don't survive contact with other people
Notes are written for an audience of one — future-you, in about eight hours, who already has all the context loaded. A status update is written for an audience of ten who have none of it. Those are genuinely different documents, and the gap between them is exactly the ten minutes I was spending every morning, badly, under deadline pressure. I'd either over-explain (the full courtroom drama of the race condition) or under-explain (it's fixed, moving on) and neither one actually told the team what they needed: is this blocking anyone, and what should I expect from you today?
So I wrote the Turn Messy Notes into a Clean Status Update prompt and pointed it at Claude Sonnet 4.6. I picked Sonnet on purpose. It's fast enough to run while the call is still connecting, and it's genuinely careful about not adding things that aren't in the source, which turned out to matter far more than I expected. I paste my notes, set the audience to 'my engineering team', set the format to 'Slack', cap it at thirty words, and hit go.
Out comes something like: 'Fixed a race condition in auth middleware — temporary sleep in place, proper fix tracked in #1847. Following up with Raj on the DB migration ETA.' Twenty-eight words. It reads like a competent adult wrote it after a full night's sleep. Nobody's eyes glaze. We move on to the next person, and I have not aged visibly during my turn.
The one constraint that earned its keep in week one
The single most valuable line in the prompt is the one telling Claude to never fabricate a date or ETA, and to write [CONFIRM: ...] instead. On day two it caught me cold. My notes said 'migration soon' and the model, rather than inventing 'migration shipping Friday,' returned 'DB migration [CONFIRM: is there a target date?]'. There was no Friday. I had no idea where my brain would have pulled 'Friday' from, but I absolutely would have said it, and then I'd have spent the following week explaining to the team why the Friday I promised had quietly become a Wednesday three weeks later.
That little bracket has saved me from at least three confidently-wrong updates since. It turns out a huge fraction of bad status updates aren't lies — they're the speaker's brain helpfully autocompleting a detail under pressure. Sonnet simply refuses to autocomplete, and asks me instead. I'd take that trade in any tool.
- Paste notes, set audience + format + word limit, done in under thirty seconds
- The [CONFIRM: ...] flag stops it from inventing dates, ETAs, or names you never wrote
- Same prompt handles Slack one-liners, email recaps, and 3-bullet standup formats
- Capping the word limit forces it to keep only what actually matters
- Setting the audience flips the register — peers get casual, leadership gets crisp
The compounding return on a thirty-second habit
It's a tiny thing. But tiny things you do every single working day compound in a way that's easy to underestimate. I figure I've clawed back three or four hours a month of actual time, and — more importantly — I've stopped starting every morning with a small spike of dread. The standup went from a thing I braced for to a thing I barely think about, which is the highest compliment you can pay any piece of a workday.
When the update is messier than usual — a half-finished investigation with three open threads and no clean conclusion — I'll run the output through my Tighten a Wordy Paragraph to Half Its Length prompt afterward to squeeze it down even further before it goes in the channel. Grab the Turn Messy Notes into a Clean Status Update prompt on Prompt Dock and run it on your own pile of 11pm shorthand. The first time it flags a [CONFIRM] you didn't realize you'd glossed over, you'll keep it in your morning routine forever.