Core Infrastructure _

Protocol

The absolute base layer for autonomous AI interactions. Bouclier enforces strict on-chain limits using composable policy contracts and a decentralized verification network.

Design Principles

Immutable Guardrails

Policy contracts are deployed as immutable bytecode. Once set, not even the deployer can modify boundaries.

Composable Policies

Stack multiple policy modules per agent — transfer limits, scope restrictions, rate limiters — all composable.

Sub-200ms Verification

Sentinel nodes verify agent actions against policy constraints before settlement, targeting sub-200ms latency.

Full Auditability

Every policy check result is recorded on-chain. Complete, tamper-proof audit trails for every agent action.

Permissionless

Anyone can register an agent, deploy a policy, or run a sentinel node. No centralized gatekeepers.

ZK-Ready Architecture

Protocol is designed for future ZK proof integration — prove policy adherence without revealing proprietary models.

Architecture Stack
Application Layer
AI Agents (LangChain, AutoGPT, Custom)
Verification Layer
Bouclier SDK + Sentinel Network
Policy Layer
IBouclierPolicy Composable Modules
Registry Layer
Agent IDs, Policy Bindings, Revocation
Settlement Layer
Base L2 — Immutable On-Chain State
Policy Interface
IBouclierPolicy.sol
// SPDX-License-Identifier: MIT
pragma solidity ^0.8.24;

interface IBouclierPolicy {
    /// @notice Validate an agent action against this policy
    /// @param agentId The registered agent identifier  
    /// @param target The contract being called
    /// @param value The ETH value being sent
    /// @param data The calldata being executed
    /// @return valid Whether the action is permitted
    function validate(
        bytes32 agentId,
        address target,
        uint256 value,
        bytes calldata data
    ) external view returns (bool valid);
}

Explore the Protocol

Read the technical documentation or start building with the SDK.

[SYS] BOUCLIER.ETH v0.1.0-alpha