The Fine Art of Frustrating the Adversary

* Eight Cisco Talos researchers shared methods to frustrate adversaries.
* Deception techniques include honeypot accounts, false infrastructure, and tarpits.
* Behavioral detections involve tighter control of legitimate remote-management tools and clear boundaries around AI agents.
* Breaking dependencies between stages of an operation prevents an adversary from reaching the next objective.
* Organizations can restrict access to critical servers based on credentials and monitor administrative changes.
* Creating honeypot email accounts using leaked domains can provide early tactical intelligence.
* False servers, shares, user accounts, and network space can be used for deception.
* Adversaries can be slowed down by throwing incoherent text at scrapers in the tarpit space to slow AI systems.
* Behavioral detections require identifying adversary techniques, potential procedures, consistent behaviors across changes, and accounting for obfuscation.
* Application allowlisting can restrict access to remote monitoring and management tools.
* Each AI agent session should have a unique identity, short-lived credentials, and traffic routed through an independent gateway.

The core pattern observed is the shift from reactive defense based on blocking known indicators to proactive defense focused on increasing the cost and uncertainty of the adversary’s decision-making process. This operates by exploiting the asymmetry between the adversary's need for speed and the defender's ability to impose friction. The emphasis on dependency breaking—forcing an adversary to re-establish infrastructure or change routes—suggests a systemic view where success is measured not just in stopping a specific intrusion, but in disrupting the entire operational tempo. Deception techniques serve as an information generation layer, pushing the adversary into environments where their expected results are unreliable, which directly feeds the behavioral detection mandate. This implies that effective defense requires shifting focus from merely recognizing malicious artifacts to engineering systemic friction points across infrastructure, identity, and process flow. The implication for agency is that resilience comes from making subsequent actions more complex, rather than simply hardening existing perimeters.
Bridge Questions: If frustration is the goal, where does the organization risk over-constraining legitimate operational necessities? How do organizations assess the long-term cost of increased complexity versus immediate risk reduction? What are the systemic consequences when successful disruption forces an adversary to abandon a strategy entirely?