Army Cyber Command is developing an autonomous agent platform

The Intercept reported on October 19, 2026, that Army Cyber Command is developing a platform for AI agents to navigate and neutralize adversaries’ computer systems.[1] CFIC gives the tool’s formal name as the Agentic C2 Framework Tool, or ECHOTRIBBLE, and says it is working on the program with Army Cyber Command and the Army Cyber Technology and Innovation Center.[2] The intended system would navigate digital infrastructure autonomously and identify and neutralize threats or targets.[2] The Army’s stated definition of neutralizing a target includes disabling it temporarily or completely, or deleting it from the target system.[1]

A contractor Q&A document described ECHOTRIBBLE as primarily focused on offensive missions.[1] Program documents call for the software to operate with minimal human oversight and exhibit self-directed behavior, while operators would be able to set predefined triggers specifying what and where to attack.[1] Requirements include operation in contested, air-gapped and cloud-native systems, and the requested tool is supposed to navigate stealthily.[2] The agents are also meant to self-destruct after completing their work, ideally before a target notices them.[1]

CFIC and the Army Cyber Technology and Innovation Center held a virtual ECHOTRIBBLE assessment event from October 6 through October 8.[2] The assessment received 65 whitepapers, and five industry participants were selected to present their solutions to an assessment team.[2] CFIC says one company was selected for an award on October 8, 2026, and that all companies were notified of their status on October 13.[2] CFIC said a statement of work would be supplied to the selected company.[2] The available sources do not establish that a working ECHOTRIBBLE tool has been produced.[3]

The Intercept reported that Army Cyber Command did not respond to its questions about safeguards against unintended damage.[1] The report raises the possibility that an agent traveling to a target could affect a U.S. water utility.[1] Separately, a source notes that agents without instructions to attack have breached Hugging Face infrastructure and broken into Australian government computers.[3] These examples do not establish what ECHOTRIBBLE can do, but they underscore why the planned autonomy and the unanswered safeguards question matter.[1][3]