Prompt pack · free download

The number that would prove your quarter was never collected

Paste the board export and last period’s update. Get the update, and the figures you are about to be asked for.

What you paste

sprint-14-export.csv

47 tickets closed · 132 points · 9 carried over

deploy_log.txt · 31 deploys · 2 rollbacks

spend.xlsx · 214k cloud · 38k tooling

slack thread: can we say the AI pilot saved time?

last quarter’s update, drafted, never sent

What comes back

Engineering update, Q3

Shipped. Checkout retry queue live 12 August. Failed-payment tickets down 41% in six weeks.

The gap table. Every claim that wants a figure, and what collecting it costs.

Review time saved by the AI pilotPR timestamps. Two hours, one script.
The uptime the board keeps asking aboutStatus page export. Already collected.

Predicted question. What did the 38k in tooling buy? A drafted answer is included.

‹ ›

Drag the handle. Left is the paste, right is the one page you send. Mocked up from the pack’s test fixture, so the shape is real and the numbers are sample data.

The engineering update pack is a free prompt pack that writes the engineering update for a board, an exec team or a CEO. Paste the board export, the deploy log and last period’s update into any AI assistant. You get the update, plus a table of the figures you are about to be asked for.

Who this is for

  • VPs of Engineering and CTOs who own the board or exec update.
  • Engineering managers whose update feeds a budget, headcount or date decision.
  • Anyone asked for a number they do not have, who had to answer anyway.

Before you download. This pack runs in your environment, on your data. Nothing is sent to Perform, and Perform receives no data from it at any point. It is provided free and as is, with no warranty of any kind. It is not legal, tax, security or accounting advice, not an audit and not an attestation, and no professional relationship is created by downloading or using it. You are responsible for reviewing it before you run it and for any decision you make with its output. Licensed under Apache-2.0, which includes a disclaimer of warranties and a limitation of liability.

Anything you point it at should be treated as untrusted input, because a document, a repository or a chat thread can carry instructions aimed at a model.

Version 1.0, 28 August 2026. Security contact: privacy@totalperform.com, acknowledged within five business days.

Download the prompt pack

A markdown file. Paste it into any assistant. Nothing to install.

Download the Claude Skill

The same pack, installed once and there every time.

The problem

  • You can describe the work. You cannot prove it moved anything.
  • Nobody set the measurement up, so next quarter is just as hard.
  • Tickets closed and story points go in. Nobody outside engineering can act on either.
  • The reader learns to skim, and the next update starts from behind.

What it does about it

  • Reads last period first: every promise and every gap you said you would close needs a line saying what happened.
  • Translates activity into consequence: a table of what each kind of work becomes for this audience.
  • Keeps measurement apart from inference: the figure first, your conclusion after, so a reader can argue with one and accept the other.
  • Builds the gap table: every claim that wants a figure, where it would come from, what it costs to start collecting.
  • Picks two to start: two, not all of them, each with an owner.

Two minutes to first output

  1. Gather the board export, the deploy log, spend, headcount and last update.
  2. Paste the pack and all of it in one message. Tidy is not required.
  3. Read the gap table before the update.

What it will not do

  • Invent a figure. It writes the line and marks the hole.
  • Ask you who the audience is. It states the assumption at the top so you can correct it.
  • Pretend to have a memory on a first run. It says there was no prior update.
What you pasteBoard or sprint export, deploy and incident log, spend, headcount, last period’s update
What comes backOne-page update, gap table, three predicted questions with drafted answers
TimeUnder two minutes
Where it runsYour assistant, on your machine
Tested againstA full quarter of messy inputs including a prior gap table. Four rounds of practitioner testing.

More tools and guides

Start with a conversation

30 minutes with the engineer who would do the work.