Skip to contentComing soon: SprintProof for ChatGPT and Codex

Feature · Gap review

Find the missing work before the sprint does.

“Did you think of…?” checks your plan against 153 checks written by engineers who have built these features before, and asks about what is missing. You decide what goes in.

The gap review in 21 seconds. A plan is checked against the 153 checks. One question becomes a part with its own estimate, another is marked not needed, and nothing changes without a click. Example data; the captions are part of the picture.

In short

SprintProof’s “Did you think of…?” gap review checks a feature plan against 153 expert checks in 12 areas, such as permissions, migrations, email and billing. It asks up to 8 open questions, each saying why it matters and what it usually adds. You add the answer as a part or mark it not needed. A run costs 20 credits.

Facts you can check

What you get

The questions a senior engineer would ask.

  • 153 checks in 12 areas

    Auth and permissions, APIs, data and migrations, frontend states, security, email and notifications, billing and plans, observability, accessibility, localization, performance, rollout and flags.

  • Why, not just what

    Every question says why it matters, such as “anything that emails an address a user typed can be used to send spam from your domain”.

  • The size of the answer

    Each question shows the part it usually becomes, with a range of days and a lane, for example +1–3 d · Backend.

  • Only what applies

    Checks are matched to the plan by keywords and by meaning, and the AI rules out the ones the plan already covers.

  • Your team’s own checks

    Admins add, edit and retire the team’s own checks next to the library, so last quarter’s incident becomes next quarter’s question.

  • Counted in the estimate

    Work added from the review shows up in the estimate on its own line, as “found in review”.

How it works

Run it, read it, decide.

Every AI step shows its credit cost before you click, and nothing in the plan changes until someone accepts it.

  1. 01

    Run the review.

    20 cr

    On a drafted plan, click Did you think of…?. It reads the title, the brief, “Done when”, the parts and the interview, and matches them against the library.

  2. 02

    Read the open questions.

    Up to 8, the most important first. Each says why it matters and what it usually adds, for example +1–3 d · Backend.

  3. 03

    Add it, or mark it not needed.

    Add as part puts the work into the plan, tagged “Added from review”, with its own estimate. Not needed dismisses the question. Both can be undone.

  4. 04

    Watch coverage grow.

    A coverage bar shows how much of the review is covered or answered. Run it again after big changes (20 cr); answered questions stay answered.

Why plans miss work

Most slipped dates come from work nobody wrote down: what happens to existing accounts, what a seat limit does to invites, what to do when an email bounces. Experienced engineers ask these questions out of habit. The gap review asks them every time, before the estimate is made.

How checks are matched

Every check in the library has a trigger, such as “the feature sends email to an address a user typed”, the question, why it matters, and the part it usually becomes. The review scores each check against the plan by its keywords and by meaning. The AI then reads the best candidates and drops those the plan already covers or that don’t apply.

Checks are marked must, should or consider. That sets the order you see them in.

The check library

Anyone in the workspace can browse the library by area and read each check in full. Library checks are read-only. Your team’s own checks sit next to them and are matched the same way. The planning checklist shows the kind of questions the library asks.

FAQ

Gap review: common questions

What does the gap review check?
It checks the plan against 153 checks in 12 areas: auth and permissions, APIs, data and migrations, frontend states, security, email and notifications, billing and plans, observability, accessibility, localization, performance, and rollout and flags.
How many questions will I get?
Up to 8 open questions per run, the most important first. Checks the plan already covers or that do not apply are left out.
Does the review change my plan?
No. It only asks. A question becomes a part when someone clicks Add as part, and that can be undone.
Can we add our own checks?
Yes. Admins add, edit and retire the team’s own checks next to the built-in library. The review matches them the same way as library checks.
What does a gap review cost?
20 credits per run, charged only if the review succeeds. Answering the questions is free. Credits come with the Team plan; see the pricing page.

Team

Plan the next feature before the sprint starts.

Create your account and name your workspace. Your first feature plan is free. To keep planning, subscribe to Team: the 14-day money-back guarantee applies.

Want to hear when the ChatGPT and Codex plugin ships? Leave your email.

Team size (optional)
One email when it ships. No spam.