Talk:EICAR test file

Page contents not supported in other languages.
From Wikipedia, the free encyclopedia

Potential Copyright Problem?[edit]

I am wondering whether it is actually ok to include the file itself here? isn't it copyrighted by eicar or whoever created it? or is too short to be copyrightable at all? i would be very happy if anyone could shed some light on this...User:RobertLemmen —Preceding unsigned comment added by 93.97.72.81 (talk) 20:03, 19 March 2010 (UTC)[reply]

"You are encouraged to make use of the EICAR test file" according to EICAR's page about the test file. Or are you referring to WMF policy that is intentionally stricter than copyright, like the concept of a "not free enough" license on Commons? --Damian Yerrick (talk | stalk) 14:13, 31 October 2011 (UTC)[reply]

Confirmed successes and failures[edit]

Please give the date (preferably in a signature) and nature of the scan (i.e. plaintext, encrypted, hdd, email, ...). Be sure to fully update your software so that your signed date is accurate.

Doesn't this violate Wikipedia's policy on No Original Research? --DragonHawk 20:39, 1 July 2006 (UTC)[reply]
Yes, but not badly, and at least the results are verifiable. Btw, "McAfee Total Protection for Small Businesses" obliterated the thing without giving me a chance to tell it not to. Stevage 13:48, 9 August 2006 (UTC)[reply]

Successes[edit]

Failures[edit]

Why do they detect it?[edit]

The one thing this article doesn't tell me is *why* AV programs pick it up. Are they specially taught to recognise this string, or is there something inherently "virus-like" in the file? When you think about it, since the file *is* benign, there's no particular reason they should do anything about it. It's like calling the fire brigade to tell them you're about to set off the fire alarm, then complaining when they don't show up. Stevage 13:48, 9 August 2006 (UTC)[reply]

It's more like holding fire drills, or testing your smoke alarm -- without starting a fire. Detecting the file means that your virus scanner is probably working properly. Furthermore, there's the possibility of testing your virus scanner's capability to detect viruses in compressed files, for example, without having to get an actual virus and then get your virus scanner to check it. You may want to use the Wikipedia:Reference desk for questions like this in future. Dysprosia 07:21, 10 August 2006 (UTC)[reply]
Reference desk? No, the article should explain more about how it works. heqs 14:33, 23 October 2006 (UTC)[reply]
The above guy has a point. If he still had to ask questions about basic information then this article needs expanding. The Kinslayer 12:26, 10 January 2007 (UTC)[reply]
The above point is still true today; it's not at all clear why this would be picked up as a virus by any scanning software. Tempshill (talk) 22:51, 6 April 2009 (UTC)[reply]
No, there is nothing *inherently* 'virus-like' about these files - nor is there about real viruses; this is why virus scanners must constantly update to keep up with new viruses. Virus scanners work, essentially, by matching patterns of bytes found in files on the user's machine with patterns of bytes found in their database of *known* viruses. When new viruses are found, new patterns are added to the database and this is pushed out as an automatic update to the virus scanning software. This file is detected as a virus (even though it isn't) because the virus software vendors have added its pattern to their database.62.232.250.50 (talk) 11:35, 9 November 2011 (UTC)[reply]

It makes use of self-modifying code[edit]

It makes use of self-modifying code to work around technical issues that this constraint makes on the execution of the test string.

What does this sentance mean? Regards, Ben Aveling 05:17, 28 November 2006 (UTC)[reply]

See: Self Modifying Code62.232.250.50 (talk) 12:11, 9 November 2011 (UTC)[reply]

I have restored this sentence, because it was very significative to my understanding of the EICAR file itself.

The executable code should end with two assembly language instructions: INT 0x21, INT 0x20, which respectively means "print that string" and "end". However, the instruction code for INT is 0xCD, which is not an ASCII character available from keyboard (though it is printable).

The actual file ends with the four-byte instruction pair DEC AX; SUB CX,[BX+SI+2A], which is responsible for the H+H* substring.

The EICAR file overwrites these last four bytes, in order to get the 0x21, 0x20 sequence. Most of the previous code (before the string offset) is only purposed for modifying this bytes (and passing the parameters to INT 21h).


Initial condition: AX=0 (I don't know why... it seems to be a precondition) 

CS:0101  354F21        XOR     AX,214F    //AX=214F
CS:0104  50            PUSH    AX          
CS:0105  254041        AND     AX,4140    //AX=0140
CS:0108  50            PUSH    AX
CS:0109  5B            POP     BX         //BX<-AX

Now we have BX=0140 

CS:010A  345C          XOR     AL,5C       
CS:010C  50            PUSH    AX
CS:010D  5A            POP     DX
CS:010E  58            POP     AX
CS:010F  353428        XOR     AX,2834
CS:0112  50            PUSH    AX
CS:0113  5E            POP     SI

At this point we have SI = 097B (the code is weird because the keyboard-ASCII requirement). 

CS:0114  2937          SUB     [BX],SI

word ptr [BX] contains 2B48, hence 2B48-097B = 21CD (note the bytes are reverted within the word). 

CS:0116  43            INC     BX
CS:0117  43            INC     BX
CS:0118  2937          SUB     [BX],SI

Increment BX by 2, and repeat the trick. Now word ptr [BX] contains 2A48, hence 2B48-097B = 20CD. 

CS:011A  7D24          JGE     0140

This condition always evaluates true. The string EICAR-STANDARD-ANTIVIRUS-TEST-FILE!$ is stored in CS:011B through CS:0139. 

CS:0140  CD21          INT     21
CS:0142  CD20          INT     20

Voilà!

Rjgodoy (talk) 09:48, 25 March 2008 (UTC)[reply]

If it is not clear enough I could try to explain it better... Rjgodoy (talk) 09:54, 25 March 2008 (UTC)[reply]

Update to the Adoption section[edit]

The Adoption section says that one Antivirus/Antimalware/EDR product doesn't detect it - Malwarebytes. However, it does seem to both detect and block the file, even if it doesn't show on VirusTotal

2603:6011:9403:B900:C917:690F:57A0:8737 (talk) 12:02, 10 May 2022 (UTC)[reply]

An updated version for non-16-bit-compatible NTVDM? Similar examples?[edit]

EICAR ever make one for newer OSes? If so, can someone from EICAR add it to the article? Interestingly, there's a malware called "Emptyspace" that is a text file that looks blank but is in fact executable. Not sure if that should be mentioned as it uses a similar technique. 74.196.181.80 (talk) 16:57, 12 February 2024 (UTC)[reply]

Retrieved from "https://en.wikipedia.org/w/index.php?title=Talk:EICAR_test_file&oldid=1206612881"