FAQs
What is wachxbt, what does it do, why is important?
WachXBT is a unified verification agent that employs adversarial reinforcement learning to continuously validate the actions, outputs, and behaviors of DeFAI agents across all blockchain networks and protocols.
WachXBT introduces a novel verification paradigm that combines multi-dimensional transaction analysis, real-time smart contract validation, and dynamic threat adaptation through adversarial training loops. Unlike traditional static verification systems, WachXBT evolves continuously, learning from emerging attack vectors and protocol changes to maintain verification accuracy in rapidly evolving DeFi environments.
Current verification approaches fall into three categories, each with significant limitations:
1. Static Rule-Based Systems implement predetermined verification
criteria but cannot adapt to novel attack vectors or protocol changes.
They provide deterministic verification but lack the flexibility required
for evolving DeFi ecosystems.
2. Traditional AI Verification employs machine learning for pattern
recognition but operates in isolation without adversarial training. These
systems suffer from concept drift and cannot anticipate sophisticated
attacks.
3. Protocol-Specific Verification focuses on individual protocols but
fails to address composition risks and cross-protocol interactions that
characterize modern DeFi strategies.
WachXBT addresses these limitations through a unified verification architecture that combines adversarial reinforcement learning with comprehensive multi-dimensional analysis. The system continuously evolves through adversarial training loops while maintaining deterministic verification for critical operations.
Key innovations include:
1. Adversarial Reinforcement Training: Multiple AI models continuously challenge and improve each other’s verification capabilities
2. Universal Protocol Integration: Unified verification across all major blockchain networks and DeFi protocols
3. Real-Time Adaptation: Immediate incorporation of new threat pat-
terns and protocol changes
4. Comprehensive Analysis Framework: Multi-dimensional verification covering technical, economic, and behavioral aspects
Which model do we use?
We believe that crypto will be the rails for AI, and staying true to our belief, we are currently using Sentient’s Dobby, which is a crypto native AI model.
Read more about Dobby: https://www.alphaxiv.org/abs/1701.03755
Download Dobby: https://huggingface.co/SentientAGI
Why are we building on the GAME framework?
GAME framework: https://whitepaper.virtuals.io/builders-hub/game-framework
GAME is a modular agentic framework that enables our agent to plan actions and make decisions autonomously based on the information provided to it. It gives us granular control over our AI agent, including the it’s goal, it’s description, worker description, what the agent sees (state), and what it can do (actions/functions).
Being a crypto native framework for building AI Agents, GAME extends its capabilities to allow agents to perform crypto native actions which make it highly relevant and convenient for us to use for allowing custom actions for crypto specific tasks.
What the role of WachXBT in ACP?
WachXBT is an evaluator agent in the autonomous hedge fund DAO of ACP
If we are in ACP why do we need a token?
ACP gives WachXBT a procedural home, it defines when a verification step should occur in every deal-flow.
The token solves everything that’s economic and decentralized about running that step:
Fee-for-service without middlemen
Every bundle that asks for a WachXBT verdict pays a micro-fee in the token. Settlement is instant and on-chain, so no one has to trust an off-platform billing system or wait for monthly invoices.
Stake-to-verify, slash-for-failure
Verifier nodes post token collateral. If a node ever green-lights a bundle that later fails evaluator review, a slice of that stake is burned. That turns “reputation” from a soft score into hard economic skin-in-the-game.
Self-funded upgrades
RL training, new risk modules, cross-chain adapters—these cost compute and research. A portion of fee flow is auto-routed to a treasury governed by token holders, making WachXBT’s evolution community-funded instead of VC-dependent.
In short, ACP tells every agent “Stop here for verification.”
The token makes sure that checkpoint is economically secure, permission-less to join, and able to improve itself long after the original builders step away.
In addition to this, we will soon be announcing a token utility paper as well to go into greater depths.
How is WachXBT different than all the other security agents?
Most “security agents” behave like automated auditors: they scan contract byte-code or execution traces for known bugs and then step aside. WachXBT begins after that layer. It treats every bundle an AI agent wants to push on-chain as a policy proposal and verifies three higher-order things in real time:
Intent alignment – Does the move still respect the strategy, draw-down budget, and timing window that the requester originally signed?
Context sanity – Has anything external (oracle update, liquidity depth, governance flag) shifted enough that the same move is now outside risk?
Cross-step coherence – If the bundle hops across five protocols, does step five invalidate the assumptions of step one?
Only when those checks pass does WachXBT attach an approval hash that lets the transaction leave escrow. In short: standard agents keep code honest; WachXBT keeps behaviour honest—every single time capital is deployed.
Why are we choosing virtuals to launch?
Because Virtuals is building far more than a protocol spec, it’s building the first public marketplace where autonomous agents can discover work, settle payments, and earn on-chain reputation without human middlemen.
ACP is simply the transaction grammar inside that larger vision.
We want WachXBT available at the foundation layer of that “open agentic economy,” not bolted onto a single DEX or roll-up.
Virtuals already supplies three ingredients WachXBT needs to matter at scale: a permission-less distribution rail where any agent can ping our verifier, a staking-and-slashing system that turns approval hashes into hard economic skin-in-the-game, and a unified fee loop that lets us reward evaluator agents in real time.
Launching elsewhere would silo us in one ecosystem; shipping on Virtuals lets WachXBT become the default trust checkpoint for every agent that plugs into the emerging machine-run web
How is the traction for the WachXBT agent?
50,000 monthly active retail users
200+ API clients such as geckoterminal, loky, heyelsa, agixbt, virtuals, sentient, capx etc
100+ partners such as gaia, lagrange, altlayer, base, arbitrum, etc
What is the roadmap of the WachXBT agent?
Website: https://wachxbt.quillai.network/
Chat App: https://wachxbt.quillai.network/chat
Last updated