Home / Blog / Hidden overlays
Detection

Hidden Overlays: How Invisible Windows Beat Screen Sharing

A screen share is supposed to be ground truth, what the interviewer sees is what's on the candidate's screen. Hidden-overlay tools break that assumption by rendering a window the candidate can read, but that screen-capture quietly leaves out.

How a window can be visible yet uncaptured

Modern operating systems let an application opt a window out of screen capture. The feature exists for good reasons, hiding a password manager or DRM-protected video from a recording. The same flag, applied to an answer panel, produces a window the candidate sees in front of them while the shared feed shows only the interview app behind it.

WHAT THE CANDIDATE SEES interview call answer overlay visible to candidate WHAT SCREEN-SHARE CAPTURES interview call overlay omitted
The overlay is fully visible to the candidate (left) but excluded from the shared feed (right).

Why you can't rely on the video feed

If the cheat is invisible in the very channel you're watching, watching harder won't help. The interviewer can stare at a clean screen share for an hour and miss an overlay feeding answers the entire time. Detection has to come from the operating system, not the video.

How hidden overlays are detected

  • Window enumeration. The agent on the machine can see windows the capture stream cannot, including those flagged to exclude from capture.
  • Capture-exclusion flags. A window explicitly marked invisible-to-capture during an interview is a strong, specific signal.
  • Z-order and focus. A top-most window receiving focus that never appears in the share corroborates the finding.
  • Correlation. Pair the overlay's presence with answer timing for a near-certain conclusion, as covered in detecting AI assistance.

Why overlays are especially dangerous in technical interviews

Overlay tools are effective because they fit the pace of live problem solving. A candidate can keep the interview window visible, maintain eye contact on camera, and still read a hidden answer panel sitting over their IDE or browser. That makes overlays more deceptive than an obvious tab switch, because the interviewer sees continuity while the candidate sees assistance.

What interviewers usually miss

Teams often assume that "share your screen" is a meaningful integrity control. It is not, at least not by itself. A candidate can comply fully with a screen-share request while still using a capture-excluded overlay. The issue is not interviewer attentiveness. The issue is that the capture channel is blind to the thing you are asking it to prove.

How to respond without overreacting

A hidden overlay signal should trigger review, not improvisation. The best process is to let the interview finish, review the evidence trail, and compare the signal timing against the answers that were given. If the overlay appeared only briefly and without other corroboration, it may be a benign application. If it stayed active during question prompts or answer formation, the finding becomes much stronger.

Key takeaways

  • Some windows are deliberately excluded from screen capture by design.
  • You cannot catch in the video feed something the feed is built to omit.
  • Detection must observe the OS window state on the machine itself.
  • A capture-excluded window during an interview is a high-severity signal.

Why screen share creates a false sense of safety

Interviewers often assume that if they can see the candidate's shared screen, they can see the candidate's working environment. That assumption is no longer reliable. Modern operating systems and desktop applications can mark windows so they are visible locally but excluded from capture APIs. The interviewer sees a clean IDE or browser. The candidate may also see an answer panel, transcript, or prompt window layered above it.

This is why hidden overlays are so effective in live technical interviews. The candidate can maintain visual continuity, avoid obvious tab switching, and read assistance while the shared screen looks normal. Human observation alone has very little chance of catching that pattern unless the candidate makes a separate behavioral mistake.

Evidence that distinguishes overlays from normal windows

A fair detector does not punish every floating window. Many legitimate tools use overlays: accessibility tools, password managers, video-call controls, note utilities, or operating-system notifications. The risk comes from a combination of capture-exclusion behavior, window timing, known process families, foreground relationship to the interview task, and whether the overlay appears during answer generation moments.

The report should therefore show the window class, capture behavior, process context, and timeline. A hidden overlay that appears for the entire interview and sits over the coding workspace deserves a different review than a normal notification that appears for two seconds.

How to reduce false positives

  • Whitelist common accessibility and meeting-control overlays where their behavior is expected.
  • Distinguish capture-excluded windows from ordinary topmost windows.
  • Score overlay events higher when they coincide with question answering, paste bursts, or AI-tool process activity.
  • Show reviewers the timeline instead of hiding the evidence behind a generic warning.

What to tell candidates before the session

Hidden-overlay detection should not be a surprise. The candidate disclosure can say that the interview agent checks for windows and applications that are hidden from screen sharing or configured to evade capture. It should also say what is not collected: the system does not record the screen, read window contents, capture audio, or log keystrokes.

This disclosure helps honest candidates clean up their workspace before the interview and reduces disputes later. If a candidate uses accessibility tooling or an approved overlay, they can disclose it. If a prohibited answer overlay appears during a closed-book round, the reviewer can point back to a rule the candidate saw in advance.

Transparency does not weaken detection. It improves fairness and makes the evidence easier to use when a real violation occurs.

Why this belongs in candidate policy

Policies often mention AI tools but forget overlays. That leaves a loophole: a candidate can claim they did not "open ChatGPT" while still using a purpose-built answer layer. Candidate instructions should prohibit hidden answer displays, capture-excluded windows, undisclosed transcription tools, and remote coaching interfaces during closed-book interviews.

Naming the method makes enforcement clearer. It also gives honest candidates a simple rule: if a tool is visible only to you and helps generate or display answers during a restricted round, ask before using it.

Catch what the screen share can't

InterviewWatch flags hidden-overlay windows the moment they appear, invisible to capture, visible to your reviewer.

Try nowContact us