Visual evidence and regressions

The idea

A visible change needs captures, and before-and-after pairs where asked. A service change needs a live check. A comparison fails on anything the change broke against the base.

The problem
Visible changes were called done without anyone looking at them.
Why it matters
Each visible change has captures, and regressions against the base fail verification.
Where it stands
In production as of

A step-by-step replay of this idea is not published yet. The library adds one when a recorded run shows it.

  • Native change-proof engine

    One engine finds the change kinds from the diff and the outputs. Each kind needs its own typed proof: a UI recording, a real HTTP exchange, tests, a query, or exact document bytes. Only a person can waive a check, and a waived check shows as not passed.

  • Harness tools that run checks and keep evidence

    QA test cases run in a headless browser on the item's own previews, with screenshots. Runbooks run as drills with approval and rollback. Device flashes are checked by a read-back hash.

  • A receipt for every task

    Every finished task gets a sealed receipt: what was done, which checks ran and how they ended, the time and the cost, built from records the platform already keeps.

    Recorded walkthrough

  • Completion truth gates

    New tests must run here or have a trace for each assertion. Claimed behaviour must be tested, and the same bug elsewhere fixed or ruled out. Planned changes must exist, steps tick only on evidence, and spend over budget needs a recorded decision.

Turn the numbers into shipped work.

Crewshal runs these choices for you: a persistent AI worker with memory and rules, on your Claude and Codex subscriptions.