Adversarial Explorer
Understand how attackers think, how Web4 defends itself, and which coherence domains detect which threats. Attack patterns explained through narrative.
Why document attacks? Security through obscurity doesn't work. By explaining attack patterns openly, we invite scrutiny, improve defenses, and help humans develop intuition about trust dynamics.
7 ATTACK PATTERNS
The Legion
Basic Sybil Attack
Create multiple fake presences controlled by one adversary to artificially inflate reputation or voting power.
The Patient Infiltrator
Long-Con Attack
Build genuine trust over 100+ cycles, then exploit it catastrophically during a brief window.
The Circle of Friends
Collusion Ring
N agents mutually endorse each other to inflate all members' trust scores.
The Miser
ATP Hoarding
Accumulate ATP without spending, creating artificial scarcity.
The Arsonist
Trust Nihilism
Systematically destroy all trust relationships through mass false accusations.
The Domino Pusher
Cascade Triggering
Identify and compromise critical network nodes to trigger cascading trust collapse.
The Noise Maker
Decoherence Injection
Inject uncorrelated noise to break phase alignment between agents.
The Legion
Basic Sybil Attack
DETECTED BY COHERENCE DOMAINS
Imagine one person wearing a hundred different masks, each pretending to be a different member of the community. They vote for themselves, endorse themselves, and create the illusion of consensus where none exists.
In Web4, this is the "Sybil attack" - named after a woman with multiple personalities. The attacker creates many LCT presences, each appearing independent but secretly controlled by the same mind.
Why it's dangerous: Democracies assume one person = one vote. If someone has a hundred votes, they can manipulate any decision.
Why Web4 resists it: Each LCT must be tied to physical hardware. Creating 100 presences means owning 100 devices. The attack scales linearly with cost while detection scales superlinearly with sophistication.
Why 9-Domain Coherence Matters
Every attack creates incoherence somewhere. The 9-domain framework provides overlapping detection - an attack that evades one domain is caught by another.
Physical
Social
Economic
Attention
Trust
Narrative
Temporal
Identity
Context
Open Research Questions
- •How do patient adversaries (100+ cycle investment) behave in real deployments?
- •Can AI-assisted attackers coordinate more effectively than human collusion rings?
- •What's the minimum coherence disruption needed to cascade into system failure?
- •Are there novel attack patterns we haven't anticipated?
If you discover attack patterns we haven't documented, please report them responsibly.