# Start
* Taking into consideration 3-weeks sprint duration
* Unit and integration tests
* Start adding independent stories with clear acceptance criteria
* Communicate better between DEV/QA so we don't get so out of sync (eg: QA testing stuff that is out of sprint or from last sprint)
* Fix the Gitlab bot so it does not spam Jira with comments
* its better to include estimations for miscellaneous tasks(for setup issues/devops/jenkins jobs updation)
* Improve our dev-deployment flow to reduce friction between teams & be able to deploy more often
* Involving QA in requirements gathering
* Consider security explicitly during the software requirements process.
* Assigning all the newly created tickets/bugs to PO/tech lead, so they triaz & decide its priority and good to plan the sprint
* Adopt Thread modeling during Design phase to detect the attack surface :)
# Stop
* Usually one person has a lot of tasks, and there is people that doesnt have any
* Accepting tickets with no translations/templates ready
* Signing contracts with third-parties before proper technical assessment
* Removing devs out of the Sprint for unplanned work
* Assigning sub-tasks directly to developers. Each developer should be responsible to choose their own subtasks, stories must continue having story leader though.
* Switching developers context, developers should focus on one epic (not story) per sprint.
* Using e-mail templates design directly on repositories. (move template design to sendgrid).
* Asking every team member on daily meetings what they done on day before, instead, look at all tasks and ask the progress of them (focus on the task).
* Generic descriptions on everything (be specific)
* Avoid pinging or assigning random dev from the team for bugfixes or tasks in the middle of the sprint, this might make us switch context and be unproductive :(
* Stop posting every minor things in public channels
# Continue
* Having the team questioning added value of features and trying to understand better business side
* Pair programming :hearts: