The Army Wants a Tool That Sends AI Agents Into Enemy Networks to Hunt on Their Own. It Is Called ECHOTRIBBLE.

Army Cyber Command has picked a company to build a tool that sends AI agents into an enemy network to find and neutralize targets on their own. Documents obtained by The Intercept, published on October 19, 2026, describe the program by its codename: ECHOTRIBBLE 1. The paperwork is not classified. It sits on the public site of the Cyber Fusion Innovation Center, the defense nonprofit that runs the Army's industry outreach 2.

The program's full name in that paperwork is the Agentic C2 Framework Tool. "C2" is command and control. The release asks industry for "a modular, autonomous agent capable of navigation, environmental observation, and anomaly detection across digital environments", built to "autonomously navigate infrastructures, identifying and neutralizing identified threats/targets" 2.

The same release sets the conditions. The tool "must operate with minimal human oversight, remain lightweight, and be deployable in contested, air-gapped, or cloud-native systems" 2. An ARCYBER document circulated to contractors goes further, requiring the software to be capable of "Self-Directed Behavior" alongside initial "predefined triggers" that set what and where to attack 1.

"Neutralize" is defined in a Q&A document for contractors as "disabling the threat completely or temporarily, deleting the threat from the target" 1.

The Intercept describes a tool that works "almost like a computer worm, silently navigating a foreign network and hopping from device to device until it finds the Army's quarry". When the job is done the agents would self-destruct, "ideally without having been noticed by the target at all" 1.

A slide deck shown to vendors at a June 24 ARCYBER event walks through one hypothetical: uploaded to a laptop, the software "identifies other potential targets" on that network and produces "a replica of itself tailored specifically for the iPad, Apple desktop, and Windows 11 PC" 1.

The money moves through a white-paper competition. The June 24 collaboration event in Augusta, Georgia drew 79 registrants and 43 attendees from 29 organizations 2. Contractors then submitted 65 white papers through the Vulcan platform; five were selected to demonstrate to an assessment team between October 6 and 8; one company was selected for award on October 8, and every company was told where it stood on October 13. A statement of work is to follow 3.

The winner has not been named and the contract's value has not been published. The same page describes the work it is paying for as supporting "forensics capabilities", "malware analysis, incident response, and support to cyber defense operations of Army networks and critical infrastructure" 3. The Intercept's reporting, based on the contractor documents, describes the program as "primarily focused" on offensive missions 1.

David Krueger, a machine-learning professor at the University of Montreal who founded the group Evitable, told The Intercept: "Given the known failure modes of the current AI paradigm, making autonomous AI systems with a mandate to 'neutralize threats' is incredibly dangerous" 1.

Hyrum Anderson, chief executive of the AI security firm Abundant and a former Microsoft and U.S. government security researcher, asked what route the tool would take to its target: "Will it go through a US water utility on its way, because that's where the adversary has staged its command and control, and take that down too?" He named the organizations at risk as the ones below "the security poverty line" — small businesses, dental practices, law firms, critical infrastructure 1.

William Marcellino, a senior behavioral scientist at RAND, put the gap this way: "One of the reasons that we trust 19-year-old Marine corporals in Ramadi with live rounds and a weapon is because they have both the compute substrate — the brain and the eyes — but the deep world knowledge to understand, 'Yeah, I probably shouldn't do that.' That is not something that current public AI technology has" 1.

The reporting sets the program against a summer of agents doing unintended things in systems they had no orders to attack: the Australian government's computers, Hugging Face's infrastructure, packages uploaded to RubyGems, an attempt on the U.S. Education Department 1. Gizmodo's write-up of the same documents notes the program is "still in the whitepaper phase" 4.

ARCYBER did not respond to The Intercept's questions, including on what would stop the tool causing inadvertent damage 1.

"ECHOTRIBBLE" is an original New York drill song and music video built from those documents — one song, one day's news.