The delivery problem for an attacker is not writing the payload. It is getting someone to run it. On streaming sites that problem is solved by context: a visitor who is already trying to make a video play will accept an explanation for why it is not playing.

Nearly every disguise below is a variation on that single technique. The countermeasure is correspondingly simple to state and hard to break: no streaming site needs you to install, extract or paste anything to play video. A modern browser decodes standard video natively. Any page saying otherwise is lying, and the specific wording of the lie does not matter.

What are the common disguises?

Seven shapes account for most of what circulates.

Disguise The claim What it actually is The tell
Missing codec "Install this codec pack to watch in HD" An installer, usually bundled adware or a loader Browsers have decoded H.264 natively for over a decade
Player update "Your player is out of date" Same, styled as the site's own UI Web pages have no player to update
Browser update A page mimicking your browser's update screen, matched to your user agent An executable Browsers update themselves and never do it from a website
Fake CAPTCHA "Verify you are human: press Win+R and paste this" You run the attacker's command yourself No verification system asks you to use a terminal
Download helper "Get the app for faster streaming" An installer with persistence The site works without it
Subtitle or archive file A .srt, .zip or .rar alongside the video A script or executable inside Extensions and archive contents rarely match the description
Signed installer A legitimately code-signed setup program Real signature, abused or stolen certificate A valid signature proves who signed, not that it is safe

The fake CAPTCHA family — sometimes described as "paste and run" — deserves special attention because it defeats file-based defences entirely. The page places a command on your clipboard and instructs you to paste it into the Run dialog or a terminal, under the pretext of verification. No file is downloaded, so download reputation checks and many scanners never engage. Nothing legitimate on the internet has ever asked you to do this.

How do file names hide what a file is?

By exploiting the fact that most systems hide extensions by default.

Trick Appears as Actually is
Double extension video.mp4.exe An executable, displayed as video.mp4 if extensions are hidden
Shortcut file movie.mkv.lnk A shortcut that runs a command
Script disguised as media episode.mp4.js A script the system will execute
Installer masquerade codec_setup.msi An installer, exactly as labelled — the label is the disguise
Right-to-left override videoexe.mp4 The characters are reversed by an invisible control character
Password-protected archive film.zip, password in the page Contents cannot be scanned until extracted

The first fix is a settings change: enable file extension display in Windows Explorer or macOS Finder. It costs nothing and neutralises half of this table permanently.

The password-protected archive is a deliberate anti-scanning measure. Any download that arrives with a password in the page text is asking you to bypass your own antivirus, and there is no legitimate reason for a public download to be encrypted this way.

How do you check a download before opening it?

A short sequence, in order of usefulness.

Check What it catches
Show extensions, then read the real one Double extensions and script files
Compare the file size to the claim A "1080p film" of 3 MB is not a film
Scan with a multi-engine service Known samples; fresh ones frequently pass
Inspect the digital signature Unsigned installers, or signatures naming an unrelated company
Note whether it wants administrator rights Media files never need elevation
Do not disable SmartScreen or Gatekeeper These warnings exist for exactly this case

The rule underneath all of it: you are not evaluating whether the file looks safe. You are evaluating whether you had any reason to download an executable at all. On a streaming page, the answer is no, and that answer is reliable in a way that scan results are not.

What should you do if you already ran something?

Assume credential theft first. Modern commodity malware is overwhelmingly information-stealing, and its first action is usually to copy browser cookies, saved passwords and session tokens — which is why changing a password from the infected machine accomplishes nothing.

Step Reason
1. Disconnect from the network Stops further exfiltration and remote instructions
2. Change passwords from a different, clean device The compromised device may be logging what you type
3. Sign out all sessions on important accounts Stolen session tokens survive a password change
4. Check for new browser extensions and profiles A common persistence mechanism
5. Check startup items and scheduled tasks The other common persistence mechanism
6. Run an offline scan from bootable media An active infection can interfere with scanners running inside the OS
7. For anything with stored payment or identity data, rebuild the system The only reliable remediation once code has run with your privileges

Step 2 is the one most often done in the wrong order, and step 3 is the one most often skipped entirely. A session token exported before you changed anything keeps working until the account explicitly invalidates it.

Related questions