Solutions / QA Engineers
For QA engineers

Test plans that already know what changed

A useful test plan is not a checklist. It is a read of the diff, the related areas of the system, the user-facing behaviour the change touches, and the things that have broken in this neighbourhood before. Most teams ship test plans that are a checklist anyway, because writing the read takes too long.

Dawn writes the read. The agent has the PR, the related tickets, the linked design, the relevant memory pages, and the connected source files. `/pr-test-plan` and `/ticket-test-plan` turn that into something a QA engineer can actually start from — and edit, instead of typing from scratch.

Pre-release pass, with help

A few moments where Dawn changes the rhythm

01

A PR is up and the change is bigger than it looks

Hand it to `/pr-test-plan`. The agent reads the diff and the related areas of the system, surfaces the user-facing behaviour to verify, calls out the regressions worth checking, and produces a plan that does not pretend the diff is smaller than it is.

02

A bug ticket lands without a clear repro

Ask Dawn to read the customer report, the linked logs, the related code path, and the related tickets, and to draft a reproduction with steps and expected vs actual. You go from "I cannot reproduce this" to "I have three things to try".

03

A release is going out tomorrow

Ask the agent for a release-pass plan: the merged tickets in the window, the user-facing changes, the historical hot spots, and the known-flaky areas. The pass starts from a real picture of what is in the build, not from last release's checklist.

04

A regression is reported on a feature you have not touched in months

Ask Dawn to walk back the recent changes that touched that area or its dependencies. The agent surfaces the candidates and the related tickets so you can narrow the bisect to a usable set.

From tickets to test plans

The Dawn pieces that QA teams reach for

/pr-test-plan and /ticket-test-plan

Coverage suggestions grounded in the actual diff or ticket — not a generic checklist.

Global command pack

Memory pages for known-flaky areas

The "this area is fragile, run the manual case" knowledge that lives in QA heads gets written down once and applied to every plan after.

Memory pages

Connected trackers and repos

GitHub, GitLab, Azure DevOps, Jira, Linear — the agent reads the same source-of-truth the rest of the team does.

Connected sources

Background tasks for big repro hunts

A broad search across logs and tickets can run as a background task while you carry on with the rest of the test pass.

Background tasks

Files browser for test artifacts

Reproduction recordings, generated test data, and HTML artifacts of test reports all live in one browseable surface, attached to the thread that produced them.

Files browser

Try it

Asks that earn their keep

Test the PR

/pr-test-plan https://github.com/example/app/pull/249

Test the ticket

/ticket-test-plan PROJ-1842

Reproduce the bug

From this customer report, draft a reproduction with steps, expected, and actual — link the relevant code paths

Plan the pass

Build a release-pass plan for tomorrow's deploy: merged tickets in the window, user-facing changes, historical hot spots

Walk back the regression

Tell me what has changed in the past 30 days that could have caused a regression in the export PDF flow

A QA-specific tip

Memory pages beat checklists

A test checklist goes stale within a quarter. A memory page that says "this area depends on these three subsystems and has the following historical regressions" stays useful, because Dawn applies it to every new test plan automatically. Migrate the wisdom out of checklists and into memory.

Memory pages in Dawn

Start asking your system questions.

Enter Dawn, connect your tools, and get your first grounded answer in minutes.

Live now. Built for real teams and real systems.