Skip to content
See the World Through ScienceA project of ALLATRA

Symantec Says the Attack Code It Found at an African Stock Exchange Was Likely Written by AI

AI & Technology

Republish this story

Our work is licensed under Creative Commons BY-NC 4.0. You may republish this piece for free — with credit to ALLATRA Media and a link to the original, unedited beyond length trims, and not for commercial use.

Read the full license

Signage reading Symantec World Headquarters, 350-380 Ellis Street, on a lawn in front of low glass office buildings.
The Ellis Street campus in California, photographed April 7, 2013. The toolkit was analyzed by this company and its Threat Hunter Team published the findings (illustrative)."Symantec Headquarters Mountain View" by LPS.1, via Wikimedia Commons, CC0

Symantec's Threat Hunter Team says a 41-module PowerShell toolkit recovered from a compromised server at a stock exchange in an African country was probably written, or at least assisted, by an AI model. The company published its account of the intrusion on Nov. 19, 2026, and attributes the toolkit to no group.

Symantec's case rests on how alike the modules are, and its report says that likeness is also what gives defenders something to look for: one common scaffold, repeated in all 41 scripts.

Every module opened with the same error-handling line and carried its own copy of the routine that sent results out as domain-name lookups rather than web requests — plumbing a programmer would normally write once and call from each script. One module held a counter none of the others had, which the report reads as a single script revised on its own. The report says none of the signs is conclusive by itself, and that Symantec could not identify which model may have been used.

The modules ran on an internet-facing server on Sept. 20, 2026, in a window of three minutes and twelve seconds, a new one starting every few seconds. According to Symantec, they searched for passwords, cloud keys and database connection strings, then tested what they found against the exchange's own investment interface and the gateway of a securities depository, where share ownership is recorded. The results went to a collector the team cannot see, so whether any of the credentials or keys worked is not known, and Symantec found no sign that the attackers widened their access or took data.

Symantec attributes six implants found on the same server, which kept the attackers' access open, to the China-linked espionage group Witchetty with high confidence. It has not attributed the toolkit to any group, and calls it more advanced than tooling previously tied to Witchetty. Symantec dates the earliest evidence of malicious activity on the server to April 2026, and says the attackers may have been present for longer.

Sources

Spot an error?

Spot an error?

Report an error

Spotted a mistake on this page? Tell us what's wrong and our editors will take a look.

What kind of problem?

Only if you'd like us to be able to follow up. We won't use it for anything else.

We correct mistakes openly. Select any text to flag it. Fixes are logged under our Corrections Policy.

Report an error

Reporting on

Symantec Says the Attack Code It Found at an African Stock Exchange Was Likely Written by AI

What kind of problem?

Only if you'd like us to be able to follow up. We won't use it for anything else.

We read every report. Corrections are logged publicly.