Help center
Open dashboard

Atarim for QA Engineers: Tracking and Managing Feedback

File bugs that carry their own repro context — element, URL, browser, viewport, and a screenshot — then verify the fix in the same thread.

Atarim team Updated 6 Aug 2026 · 10 min read
Role-Specific Guides
Atarim bug tracking with reproduction context captured on the page
Before you start

Relevant for

  • QA engineers testing websites and web applications, filing defects, and verifying fixes before release.

Required knowledge

  • Standard testing practice. Familiarity with browser devtools helps but is not required.

Tools & resources needed

  • A team member role or above, access to the projects under test, and a publicly reachable environment if you want the AI diagnostics.

Most of a QA engineer’s wasted time goes on the same two things: writing down context a tool could have captured, and re-testing something that was closed without anyone telling you. Atarim removes the first by recording repro conditions when the task is created, and the second by giving you a status that means “ready for you to verify”.

Everything a developer would otherwise ask for, captured at the moment the bug was found.

Most of a QA engineer’s wasted time goes on the same two things: writing down context a tool could have captured, and re-testing something that was closed without anyone telling you.

Atarim removes the first by recording repro conditions when the task is created, and the second by giving you a status that means “ready for you to verify”. This guide is about testing and verification specifically, rather than general task use.

Filing a Defect

The click is what carries the context — element, screen size and browser are captured with it, so the developer does not have to ask.

Instructions:

  • Open the page under test and switch to Comment mode in the bottom bar.
  • Click the element that is behaving incorrectly. A marker is placed and a comment box opens, prompting What are you thinking?
  • Describe what you did and what happened.
  • Select Comment to post. A task is created, and Atarim captures a screenshot of the page as it was — described as a snapshot of what was seen when the task was made.
  • You do not have to write down the URL, the browser, the viewport, or where on the page it happened. That is recorded for you.

    Clicking an element in Comment mode to file a defect
    Filing a defect in Comment mode
    A filed defect with its captured screenshot
    The captured screenshot
    Spend the words on what the tool cannot capture
    Since the environment is captured automatically, spend the words on the steps that produced the state. “Submit with an empty email field, then correct it and resubmit” is the half a screenshot will never show. See How To Create Task In Atarim

    What Gets Captured for You

    Expand Technical Details on any task to see the repro context.

    FieldWhy it matters in a defect report
    Element PathThe selector for the element involved, copyable in one action.
    Page URLExactly where it happened, also copyable.
    BrowserThe environment — usually the answer when a developer cannot reproduce it.
    Screen SizeThe viewport, which is the breakpoint the defect actually occurs at.
    Website Name and Page NameWhich project and page the defect belongs to.
    Created ByWho found it, for follow-up questions.
    Task ID and Share LinkA stable reference and a link for a test report or a standup.
    Due Date and Add TagsScheduling and categorisation, editable in place.
    The Technical Details panel expanded on a defect
    The Technical Details panel
    Copy the Share Link into your test run notes rather than the task title
    Anyone reading the report later lands on the defect with its screenshot and thread intact, instead of searching for something by name. Read Understanding User Roles And Permissions In Atarim

    Annotating the Evidence

    When the defect is visual, mark it up. The comment box has an annotation control, described as annotating a screenshot and attaching it as a comment, which opens a window headed Annotate screenshot.

    ToolUse it to
    ArrowPoint at the exact element that is wrong.
    RectangleBox the region where the layout breaks.
    CircleRing something without covering it.
    PenDraw freehand.
    TextLabel the expected versus actual directly on the image.
    Move/SelectReposition or resize what you have drawn.
    PanMove around a tall screenshot. Hold the space bar as a shortcut.

    Undo, Redo, and Fit to Screen are available while you work. Save attaches the annotated image to your comment; Cancel discards it.

    Use Text to write the expected result on the image itself
    A screenshot showing what is wrong is useful; one showing what it should have been closes the loop without a second message. Explore How To Use The Annotation Tools For Feedback

    Testing Across Breakpoints

    The bottom bar carries a device menu for previewing at Desktop, Tablet, and Mobile. A defect filed while a view is active is recorded against that view, so a mobile-only bug arrives already saying so.

    Marking up a screenshot with the annotation tools
    File from the breakpoint the defect occurs at — the viewport is recorded
    File responsive defects from the breakpoint they occur at
    “Broken on mobile” in a comment is a claim a developer has to reproduce; a task recorded at the mobile viewport is evidence. Learn More About Responsive Testing Across Breakpoints

    Automated Checks the AI Can Run

    Atarim’s AI team includes Glitch, which reviews functionality — interactions that break and what is breaking them. Alongside the review, the agents can run real diagnostics when you ask for them in an AI chat. These are run on request, not automatically.

    CheckWhat it returns
    Lighthouse auditPerformance, accessibility, best practices, SEO, and PWA scores out of 100, with actionable fixes.
    Accessibility scanA WCAG 2.1 pass returning violations with severity and the affected nodes.
    Colour contrast checkA WCAG 2.0/2.1 contrast ratio with AA and AAA pass or fail for normal text, large text, and UI components.
    Browser verificationA headless browser driven through a scripted sequence of interactions against a URL, returning the document status and the result of each step.
    These need a publicly reachable URL
    The Lighthouse and accessibility scans explicitly exclude authenticated and internal addresses. Staging behind basic auth, a VPN, or an IP allowlist will not return useful results.
    Treat these as a first pass, not a test plan
    An accessibility scan finds violations a scan can find. It will not tell you whether a flow makes sense with a screen reader, and the WCAG criteria that need human judgement stay yours.

    Findings from a page review come back in the Review tab in the sidebar, grouped by the viewport they were found in. Selecting one takes you to it on the page.

    Glitch has its own guide worth reading if functionality is your focus. Read Catch Bugs Before Users Do With Glitch

    Verifying Fixes

    Statuses are what turn a defect list into a verification queue.

    StatusWhat it means to you
    OpenFiled, not picked up.
    In ProgressA developer is working on it.
    Pending ReviewFixed and waiting on you. This is your queue.
    CompleteVerified and closed.

    Priorities run Low, Medium, High, and Critical, set on the same control as status.

    The device preview menu offering desktop, tablet and mobile views
    Statuses
    Do not let developers close their own fixes
    A defect moved straight to Complete has been marked fixed by the person who fixed it, with no verification step. Agree that developer-fixed work goes to Pending Review and that QA closes it — otherwise regressions ship silently.

    On comments Atarim has identified as actionable, a Show me control appears, with the tooltip describing it as showing the update in place. It previews the suggested change on the page without applying it, and carries controls to hide the preview again or regenerate it. That preview is a quick sanity check — not a substitute for retesting once a change is genuinely applied.

    The Show me control previewing a suggested change on the page
    Previewing a change with Show me

    Tracking completion honestly is what makes the numbers meaningful later. See Tracking Task Progress And Completion

    When the Element Has Moved

    A defect pinned to an element that no longer exists reports Element not found on this page, and offers Remap to element so you can point it at the current markup.

    Remapping a defect to the current element after markup changed
    Remapping a defect to the current element

    This happens constantly in an active codebase. A refactor orphans every task attached to the renamed component, and remapping is what stops your defect list quietly rotting.

    Check for orphaned tasks as part of regression testing after a refactor
    An orphaned defect is one nobody will retest, because nobody can find what it pointed at.

    Organising a Test Run

    The filters in the Tasks tab are what make a defect list workable.

    Instructions:

  • Open the Tasks tab in the sidebar. Filter and sort controls appear alongside the list.
  • Filter to Assigned To Me, or sort by Status to pull Pending Review items together.
  • Switch scope between the current page and the whole project.
  • Turn on Show Complete Tasks when auditing a release rather than working the live queue.
  • Use search to find defects and comments as you type.
  • A defect set to Pending Review awaiting verification
    A defect awaiting verification

    Tags are useful for cutting across projects — a regression suite, a release, a browser-specific set. They are added under Add Tags, where the field prompts you to type and press enter to create a new tag.

    Tag by release rather than by sprint
    A release tag stays meaningful when you are asked what shipped with a version; a sprint number stops meaning anything to anyone outside the team within a month. Explore Simplifying Task And Project Views With Atarim Filters

    Keeping Test Discussion Internal

    Tasks and comments can be marked internal, which hides them from guests and clients. Defect triage, severity arguments, and known-issue discussion belong there rather than in a thread a client is reading.

    Adding an internal note to a task
    Adding an internal note
    A task marked internal so clients and guests cannot see it
    A task marked internal
    Confirm it is enabled first
    Internal tasks depend on your plan and the control reports a lock where unavailable. A client reading your team debating whether a defect is worth fixing is an expensive way to find out. Read Internal Tasks

    Benefits for a QA Engineer

    BenefitIn practice
    No writing down the environmentBrowser, viewport, URL, and element path are captured with the defect.
    Fewer “cannot reproduce” rejectionsThe conditions are in the report rather than in your memory.
    Evidence, not descriptionAn annotated screenshot settles arguments a paragraph will not.
    A real verification queuePending Review is a status that means the work is yours.
    Defects survive refactorsRemapping keeps reports attached when markup moves.
    A cheap automated first passLighthouse, WCAG, and contrast checks before manual testing starts.

    Example Use Cases

    SituationWhat to do
    A layout bug only on mobileSwitch to the mobile view first, then file. The viewport is recorded.
    A developer cannot reproduce itPoint them at Browser and Screen Size in Technical Details.
    Accessibility pass before releaseAsk for a WCAG scan on a public URL, then test the judgement-based criteria manually.
    Verifying a batch of fixesSort by Status and work the Pending Review items.
    After a big refactorLook for orphaned tasks and remap them before regression testing.
    Reporting what shipped in a releaseFilter by release tag with Show Complete Tasks turned on.

    Known Limitations

  • AI diagnostics need a public URL. Lighthouse and accessibility scans exclude authenticated and internal addresses.
  • Diagnostics run on request in an AI chat, not automatically as part of a review.
  • Screenshots capture the moment of reporting and do not update as the page changes.
  • Element paths reflect the page as it was, so tasks need remapping after markup changes.
  • Internal tasks and attachments depend on your plan.
  • Nothing enforces who closes a defect — QA-verifies-and-closes is a team agreement, not a product rule.
  • Statuses and priorities are fixed and cannot be renamed to match another defect taxonomy.
  • FAQs

    Do I need to record the browser and screen size myself?

    No. Both are captured automatically and shown in Technical Details.

    How do I file a bug that only happens on mobile?

    Switch the device preview to Mobile first. The viewport is recorded with the task.

    Can I annotate the screenshot?

    Yes, using the annotation control on the comment box, with arrow, rectangle, circle, pen, and text tools.

    What happens to defects when the markup changes?

    They report that the element was not found and offer Remap to element.

    Which status means a fix is ready to verify?

    Pending Review. Treat that as your verification queue.

    Can Atarim run accessibility tests?

    Yes — a WCAG 2.1 scan, plus Lighthouse and contrast checks, on request in an AI chat and against publicly reachable URLs.

    Can I test a staging site behind a login?

    You can file defects on it as normal, but the AI diagnostics exclude authenticated and internal URLs.

    Can I hide defect triage from the client?

    Yes, using internal tasks and notes, where enabled on your plan.

    Can defects be pushed to Jira or ClickUp?

    Yes. Atarim integrates with the major trackers, plus Slack, webhooks, and Zapier.

    Can I rename the statuses to match our defect workflow?

    No. The four statuses and four priorities are fixed.

    Common issues

    • “Element not found on this page” — the markup changed. Use Remap to element to reattach the defect.
    • Clicking a link leaves a comment instead of navigating — you are in Comment mode. Switch to Browse, navigate, then switch back.
    • An accessibility or Lighthouse check returns nothing useful — the URL is not publicly reachable. Run it against a public environment.
    • Defects are being closed without verification — agree that fixes go to Pending Review and QA closes them.
    • Filters are missing from the sidebar — open the Tasks tab; the controls appear with it. If still absent, filtering is not enabled for your role.
    • No Show me control on a comment — it appears only on comments marked actionable, and not while a change is queued.
    • The screenshot does not match the page — it is a snapshot from when the defect was filed. The page has changed since.
    • Internal tasks are locked — not included on your plan. Do not assume a comment is hidden.
    • You cannot find a defect you filed — the list may be scoped to the current page. Widen the scope, or turn on Show Complete Tasks.
    • Technical Details is missing fields — your role does not carry full visibility. Ask an administrator.

    Conclusion

    File from the breakpoint the defect occurs at, spend your words on the steps rather than the environment, annotate when it is visual, and hold the line that QA closes defects rather than developers. The context comes free — the discipline around verification is what actually stops regressions shipping.

    It helps to understand how the developers receiving your reports work. Explore Atarim For Frontend Developers

    Tips & best practices

    • File from the breakpoint the defect occurs at — the viewport is recorded.
    • Write the steps, not the environment. The environment is captured.
    • Label expected versus actual on the annotated screenshot.
    • Copy the Share Link into test notes rather than the task title.
    • Treat Pending Review as your verification queue.
    • Never let a fix go straight to Complete without verification.
    • Check for orphaned tasks after a refactor and remap them.
    • Run automated checks first, then spend manual time on what a scan cannot judge.

    Related articles