Guide
How to use TestFlow
One pass through the whole cycle: write a test case, put it in a suite, run it, record what happened, read the report and turn each failure into a bug report. Fifteen minutes end to end, and you will have done what QA engineers do every day. New to the terms? Start with the theory articles.
1. Sign in
Use the login and password received from the learning coordinator. There is no registration page: accounts are created by the learning coordinator. If you forget your password, ask our customer support to set a new one.
A session lasts up to eight hours. When it ends you are sent back to the sign-in page; finish and save what you are editing before a long break.
Sign in2. Your workspace and projects
Everything you create belongs to you alone. Another student on the same course cannot see or change your cases, and you cannot see theirs. The same goes for your bugs. Your numbering is yours too: your TC-001 and their TC-001 are different test cases, and the same is true of BUG-1.
On the first sign-in the workspace is filled with a demo project — twelve cases about login and payments, three suites, two finished runs. It is there to be taken apart. Edit it, delete it, run it.
Work is organised in projects. The project switcher at the top of the sidebar shows the current one; switch it to see that project's cases, suites, runs and bugs. Projects are created by the teacher, and you only see the ones you have been given.
3. Write a test case
- Open Test Cases and click + New Test Case.
- Write the Title as the check itself: “Login with invalid password”, not “Login testing”.
- Pick a Type (Functional, Non Functional, Regression, Smoke) and, if it belongs to one, a Suite.
- Fill in Preconditions — what must already be true before step 1. An existing account, an item in the cart, a card that will be declined.
- Add Steps. Each step is one Action and the Expected Result of that action. Use ↑ ↓ to reorder, ✕ to delete.
- Click Save. The button turns green and confirms the save next to it.
Title and at least one Action are required — until they are filled in, Save stays disabled and tells you what is missing. When there is nothing new to save it is disabled too, so a green Save always means there are unsaved changes.
4. Group cases into suites
Open Test Suites → + New Suite and name it after a feature: Authentication, Payments, Profile. Then open the suite and use Add / Remove Test Cases to choose what belongs in it.
A case lives in one suite at a time, so moving it here removes it from wherever it was. That is deliberate: a case that belongs everywhere belongs nowhere.
5. Create a test run
- Open Test Runs → + New Test Run.
- Name it after its purpose and the build: “Regression Run #4”, “Smoke — Release 1.4”. In six weeks the name is all anyone has to tell two runs apart.
- Select the cases. Use the suite filter and the search box; the checkbox in the header selects everything currently visible.
- Click ▶ Start Run.
Choosing what goes into a run is a real decision. A smoke run is ten minutes of the most critical paths, meant to answer “is this build worth testing at all”. A regression run is everything that must not break.
6. Execute the run
The runner shows one case at a time: preconditions, then the steps with their expected results. Do what the step says, compare with the expected result, then record the outcome.
- PASS
- Everything matched. Not “close enough” — matched.
- FAIL
- The product behaved differently. A comment field appears: describe what actually happened, with the data you used. When the run is finished, this comment becomes the actual result of your bug report.
- BLOCKED
- You could not check it — environment down, feature not deployed, a dependency broken. Write why.
- SKIP
- You decided not to run it in this run.
After Pass, Blocked or Skip the runner moves to the next case by itself. After Fail it stays, so you can finish the comment. Clicking the same status again clears the result back to Not Run. You can jump around with the list on the left, and the progress counter shows how much is done.
You can close the page and come back later — the run keeps its state, and Continue in the runs list picks up where you stopped.
7. Read the report
View Report turns the run into the summary someone else can act on: totals per status, how much of the run was executed, and the pass rate among the executed cases. Below that, every case with its result and the comments you wrote on failures.
Read it in this order: what failed, what is blocked, how much is still unexecuted, and only then the pass rate. A 95% pass rate with a third of the run blocked is not good news, it is an unfinished job.
The filter next to Results narrows the table to one status — the fastest way to collect all failures.
8. Report bugs
A failed result tells you that something is wrong. A bug report tells a developer what to fix. You report bugs from the report of a finished run: once no case is left Not Run, the Bug column shows Report bug next to every failed result. Until then it says “Finish the run first”.
- Open the run's View Report and click Report bug on a failed case.
- Choose Report a new bug. The form is already filled in from what you have: the steps come from the case's actions, the expected result from its expected results, the actual result from your comment.
- Rewrite the Title as the problem, not the check: “Account locks after 4 failed attempts instead of 5”, not “Account lock test fails”.
- Set Severity (blocker, critical, major, minor, trivial) and Priority (high P0, medium P1, low P2), and add the Environment: browser, OS, build, test server.
- Click Report Bug. It gets the next number in the project (BUG-1, BUG-2…), and the number appears in the report next to the failed case.
Title, steps, actual and expected result are required. Cut the steps down to the shortest path that still shows the problem: the case may have ten steps, the bug usually needs three.
The same bug often fails more than one case, or the same case again in the next run. Do not report it twice: choose Link an existing bug and pick it from the list. Unlink in the report undoes a wrong link; the bug itself stays.
All bugs of the project are on the Bugs screen. Click a bug to read it, the pencil to edit it. A bug starts as new and is closed as fixed or rejected (not a bug, cannot reproduce, works as designed). A closed bug that comes back is moved to reopened, and from there to fixed or rejected again — check the fix in a new run before you close it.
9. Duplicate and delete
Every row in the test cases list has an ⋯ actions menu on the right. Duplicate opens a new case pre-filled with the type and suite of the original — handy for a variation like “Login with invalid email”. Delete removes the case for good.
Both ask for confirmation first, because neither can be taken back.
10. Work as a Scrum team
Everything above is yours alone. The Scrum Board is the one screen you share: every student of your cohort works on the same board, as one Scrum team. The board shows tasks of every project; each task has a Project field, filled with the project selected at the top and open to change. What you add or move, your teammates see after they reload the page — and the other way round. Your mentor sees the same board and may add tasks or sprints to it too. If the screen says you are not in a cohort, ask your teacher to add you to one.
The work goes in three steps, like a real team's sprint.
- Fill the backlog. Open Backlog and click Create New. A task gets a title, a description of at least 10 characters, a type (Epic for something a user needs, Task for technical work), a priority and a number: two letters from the project's name and a count, like DM-1, DM-2… shared by the whole team. Every new task starts in To Do; you can change its status once it is created. A task can belong to an epic: pick it in the Epic field. This is optional, and the epic's window lists its tasks.
- Plan a sprint. On the Sprints tab click Create sprint, give it a name, dates (three weeks from today by default) and a goal — one sentence on what the team wants to have by the end. Back in Backlog, click a task to read it and leave comments (up to 10 per task), press the pencil and pick the sprint in its Sprint field, estimate each in story points and pick who does it. Then pressStart sprint.
- Move the cards. The Active sprint tab shows the columns To Do,In Progress, Review and Done. Drag your card to the next column as the work moves on; on a phone, open the card and change its status. The sprint header counts how many points are done.
When the sprint ends, click Complete sprint. Done tasks stay with the sprint as its history; everything unfinished goes back to the backlog, ready for the next sprint planning. Only one sprint runs at a time on a board.
Need two sprints at once, say for two projects? Click + Team at the top of the board. A team is another board of your cohort, with its own backlog and sprints, so its sprint runs alongside the main one. The boards then show as tabs above the board — Main board and each team; a green dot means a sprint is running there. Everyone in the cohort can open and change every board. Only your teacher can delete a team. Planned a sprint on the wrong board? Open it with Edit and pick another Board: the sprint moves there with all its tasks.
To rename a sprint, change its goal or dates, or delete it, click Edit next to it — in the sprint header on Active sprint, or on Sprints for planned and completed sprints. Deleting a sprint keeps its tasks: they all go back to the backlog, done ones included.
A scale for typical tasks
Use these as anchors when you estimate: compare a new task with the closest row in the table for your role, and put it above or below that row if it feels bigger or smaller. Your team may calibrate the numbers differently — what matters is that everyone uses the same scale.
Testing (QA)
| Points | Size | Typical task |
|---|---|---|
| 1 | Trivial | Fix a typo in a test case, re-run one failed test, report an obvious bug with a clear screenshot. |
| 2 | Small | Write one simple test case, update a test case after a small change, retest a fixed bug. |
| 3 | Medium | Write a checklist or 3–5 test cases for one feature, test one screen with a form, investigate and report a bug that is hard to reproduce. |
| 5 | Large | Cover a whole feature with test cases and run them, smoke test a new build, write a test plan for a small module. |
| 8 | Very large | Test a feature that spans several screens or an API plus UI, regression testing of a release, prepare test data and an environment. |
| 13 | Too big | “Test the whole application”, or automate a full flow from scratch. Split it into smaller tasks before the sprint starts. |
Development
| Points | Size | Typical task |
|---|---|---|
| 1 | Trivial | Fix a typo or a label, change a text or a constant, bump a config value. |
| 2 | Small | Fix a simple bug with a known cause, add a field to an existing form, make a small style change. |
| 3 | Medium | Build one screen or component from a ready design, add an API endpoint with validation, fix a bug that needs some investigation. |
| 5 | Large | Implement a feature end to end (UI, API and database), integrate a third-party service, refactor a module. |
| 8 | Very large | A feature that touches several modules, a database migration with changes in the code, authentication or permissions for a whole area. |
| 13 | Too big | “Build the whole admin panel”, or rewrite a core part of the system. Split it into smaller tasks before the sprint starts. |
Design
| Points | Size | Typical task |
|---|---|---|
| 1 | Trivial | Change a colour, an icon or a text in a finished layout, export assets for developers. |
| 2 | Small | Design one state of a component (hover, error, empty), adjust spacing and alignment on one screen. |
| 3 | Medium | Design one simple screen or a form from a ready wireframe, make a screen responsive for mobile, prepare a small icon set. |
| 5 | Large | Design a feature with several screens and states, research and build a clickable prototype, create a set of UI components. |
| 8 | Very large | Design a whole user flow (for example sign-up and onboarding) with a usability check, redesign a complex screen, extend the design system. |
| 13 | Too big | “Redesign the whole product”, or create a design system from scratch. Split it into smaller tasks before the sprint starts. |
Points describe effort, complexity and uncertainty together, not only time: a short task in an area you do not know yet can be worth more than a long routine one. After a couple of sprints, use the number of points your team actually finished as your guide for how much to take into the next one.
Your turn
Write one test case, put it in a suite, run it, report what failed and read the report. That is the whole loop.