Does CodeSignal Detect Tab Switching and Copy-Paste?

CodeSignal handles the two differently. Copy and paste inside the editor is logged and shown to the employer: pastes appear as markers in the coding replay, and the Suspicion Score counts them. Tab switching is not a documented automated signal. As of September 9, 2026, CodeSignal's public documentation names no focus, visibility, or tab event; instead, leaving the test is caught by the screen recording on a proctored assessment and by the pledge you sign on any assessment or interview. So a tab switch is not counted by software, but on a proctored test it is watched by a person.

What CodeSignal logs in the editor

CodeSignal's results documentation says the employer can replay your coding session and "see if there were any suspicious paste events. They will be visible as: filterable markers, hoverable timeline events, and clickable events on the Replay Activity list."

The Suspicion Score documentation then tells reviewers how to read them. Size and count matter: "Larger chunks and/or more frequent pastes are more suspicious." Timing matters: "Pasting after having no activity in the replay for a while, and/or after a Description Copy event is a common cheating pattern." And a "Description Copy" event, meaning you copied the problem statement, is logged with a character count and position, because copying the prompt out and pasting a solution in is the pattern CodeSignal describes.

Language switches are logged too: "Switching the language combined with paste events is also a highly suspicious behavior pattern."

None of that requires proctoring. It is editor telemetry and runs on every test. The generic browser mechanics of paste events, and what they cannot prove, are in how interview websites detect copy and paste.

What CodeSignal says about tabs

The documentation approaches tabs as a rule, not a sensor.

On assessments, setup step three is "Close Tabs & Apps." The proctoring documentation says "Step 3 will prompt you to close all unrelated tabs, browser windows, and applications." The acknowledgement in CodeSignal's screenshot reads "I will not search the web; doing so will result in disqualification," while the GCA rules Q&A carves out syntax-only searches. Which version you get depends on the assessment; the one on your screen governs.

On interviews, CodeSignal's interview setup documentation describes "A pledge to close all tabs and apps aside from the interview session" with "sub-pledges that specify whether another tab can be kept open for syntax reference, whether a single monitor is required, and whether one must use only the CodeSignal IDE during the session." Employers can also "Disable description copy."

Enforcement is the screen recording. On a proctored test the display is recorded for the duration and a reviewer watches it; the documented rejection reasons include "Test-taker searched online for help (other than syntax-related searches)" and "Test-taker used an external IDE." That is how a tab switch becomes a finding: not through an event counter, but through a person seeing the other tab. What the recording covers is in whether CodeSignal records your screen.

What this means on a non-proctored test

No camera, no screen share, no documented tab sensor. CodeSignal's public documentation does not claim to detect a tab switch in that setting. What it does have is the replay: an idle gap followed by a large paste is the exact pattern the Suspicion Score guidance describes, whether or not anyone saw the tab. The browser-level events a page could in principle observe are explained in how interview websites detect tab switching; CodeSignal does not document using them.

Contrast with HackerRank

HackerRank documents tab proctoring on every test, with "Out of Window Duration" and exit counts in the report. CodeSignal documents none of that. The two platforms invest in different places: HackerRank in browser events and screenshot classification, CodeSignal in a human-reviewed recording and editor-pattern scoring. Neither reads the contents of another tab.

What this means for candidates

Treat every paste as visible with its size and timing, and every description copy as logged. Write in the editor rather than drafting outside it. On a proctored test, keep the rules literally: one window, no external IDE, searches limited to syntax only if your rules allow that. On any test, the acknowledgements you checked are the standard a reviewer applies. If the rules leave documentation or a second tab unclear, ask the requesting company in writing before you start. The full inventory is in what CodeSignal proctoring records.