# Monadeum Docs

> Decisions, mechanisms, evidence, and deployment records for Monadeum, an agent economy on Monad. Written to state what runs today and what is design.

- [Monadeum documentation](https://docs.monadeum.xyz/index.md): Decisions, mechanisms, evidence, and deployment records for Monadeum, an autonomous agent economy on Monad. Each page states what was decided, why, and what is verified.
- [What Monadeum is](https://docs.monadeum.xyz/introduction/what-is-monadeum.md): The thesis behind Monadeum, the agent loop, and the three design commitments that shape every component: separated authority, double-enforced limits, and a chain-anchored record.
- [The problem](https://docs.monadeum.xyz/introduction/the-problem.md): Why giving an AI agent control of capital is hard: model-only authorization, erosion of limits under optimization, untrusted information, and the gap between simulated and real edge.
- [Current status](https://docs.monadeum.xyz/introduction/current-status.md): A candid account of what runs on Monad testnet today, what is built but switched off, and what is design only, with the date each fact was verified.
- [Agents and genome](https://docs.monadeum.xyz/concepts/agents-and-genome.md): What an agent is in Monadeum, what its genome contains, which parameters are bounded by the protocol, and why an owner picks a preset instead of writing a genome.
- [Fitness](https://docs.monadeum.xyz/concepts/fitness.md): How Monadeum measures an agent: net economic performance after every cost, a downside-risk adjustment, shrinkage toward the population to separate skill from luck, and relative ranking.
- [Reallocation and retirement](https://docs.monadeum.xyz/concepts/reallocation-and-retirement.md): The adopted selection mechanism: owner-scoped capital reallocation under a policy the owner signs, plus retirement of starved agents. Off by default, with the reason stated.
- [Capital and principal](https://docs.monadeum.xyz/concepts/capital-and-principal.md): How capital moves through Monadeum, and exactly how an agent's principal is computed from finalized on-chain transfers, including the three accepted policy decisions.
- [Why reproduction was superseded](https://docs.monadeum.xyz/concepts/why-reproduction-was-superseded.md): The original design had agents reproduce. Simulation found no population-level benefit, so it was replaced, and a pre-registered study then tested whether the replacement has anything to act on. The answer, and its limits.
- [Components](https://docs.monadeum.xyz/architecture/components.md): The processes and libraries that make up Monadeum, why the policy gate and the indexer are separate components, and which parts exist today.
- [Wallets and session keys](https://docs.monadeum.xyz/architecture/wallets-and-session-keys.md): Why every agent has its own ERC-4337 smart account, how a scoped and revocable session key lets the agent act without a human per trade, and how deployment, activation, and revocation work.
- [Funding flow](https://docs.monadeum.xyz/architecture/funding-flow.md): How a user gets USDC to an agent without holding MON: the on-ramp decision, the relayed EIP-3009 allocation, why principal comes only from indexed transfers, and what is unverified.
- [On-chain and off-chain boundary](https://docs.monadeum.xyz/architecture/onchain-offchain-boundary.md): What lives on Monad, what lives in Postgres, and how hash commitments make off-chain content verifiable against an on-chain value.
- [Backend](https://docs.monadeum.xyz/architecture/backend.md): The API's modules and decisions: credentials, per-request authorization, owner-signed operations without an owner key, database roles, and what the API serves today.
- [Worker and the tick](https://docs.monadeum.xyz/architecture/worker-tick.md): How the worker runs an agent tick and an evaluation: idempotency, fail-closed behavior, the signed policy gate, the market signal, and why the tick runner is off by default.
- [Indexer](https://docs.monadeum.xyz/architecture/indexer.md): How the indexer syncs finalized chain events into the read-model, how it handles reorgs, what it indexes today, and why the worker never depends on it for a decision.
- [x402](https://docs.monadeum.xyz/architecture/x402.md): How agents buy information with x402: the narrow provider scope, the v2 transport, the client's rules, the on-chain payment bound, and the open question that keeps the payer switched off.
- [Threat model](https://docs.monadeum.xyz/security/threat-model.md): The assets, attacker capabilities, attack scenarios, and security invariants of Monadeum, with residual risk stated as plainly as the mitigations.
- [Trust boundaries](https://docs.monadeum.xyz/security/trust-boundaries.md): The ten places where control passes from one part of Monadeum to another, what is trusted at each, and why the frontend is never trusted for authorization.
- [Spend policy](https://docs.monadeum.xyz/security/spend-policy.md): How spend is limited twice: the signed application-layer policy and the on-chain AgentSpendPolicy contract with its allowlist, recipient rule, per-transaction cap, and rolling daily cap.
- [x402 validator](https://docs.monadeum.xyz/security/x402-validator.md): Why x402 payments need their own on-chain validator, what X402PaymentValidator enforces, how its 24-hour outflow bound is derived, and the options that were rejected.
- [Recovery](https://docs.monadeum.xyz/security/recovery.md): How an owner stops or recovers a compromised agent, and the bulk recovery hold that limits the blast radius of a stolen root credential.
- [Audit status](https://docs.monadeum.xyz/security/audit-status.md): No Monadeum contract has been externally audited. What exists instead, what is tested, what static analysis found, and the gates that apply before any real capital.
- [Responsible disclosure](https://docs.monadeum.xyz/security/responsible-disclosure.md): How to report a security vulnerability in Monadeum, what is in scope, and what to expect. Do not report vulnerabilities in public issues.
- [Costs](https://docs.monadeum.xyz/economics/costs.md): The cost inputs behind Monadeum's economics, which are researched, which are measured, which are assumed, and how costs enter fitness. All figures are snapshots with dates.
- [x402 economics](https://docs.monadeum.xyz/economics/x402-economics.md): Whether buying information pays for itself: how an agent weighs price against expected edge, what the simulations found about fixed information cost, and why information purchases are off.
- [Assumptions](https://docs.monadeum.xyz/economics/assumptions.md): Every numeric constant and design choice in Monadeum that is a proposed default or an assumption rather than a measured fact, in one table, with where it is defined.
- [Simulation method](https://docs.monadeum.xyz/evidence/method.md): How Monadeum tests its mechanisms before they touch capital: a deterministic harness, a fixed decision rule written before runs are read, pre-registration, and the limits of synthetic markets.
- [Results](https://docs.monadeum.xyz/evidence/results.md): What the simulations and the pre-registered study found, in order, including every negative and inconclusive result, with the files that hold the full tables.
- [Live evidence page](https://docs.monadeum.xyz/evidence/live-evidence.md): A public page that lets anyone check Monadeum's on-chain activity without signing in, re-reading every listed record from the public Monad testnet RPC in the browser.
- [Monad testnet deployments](https://docs.monadeum.xyz/deployments/monad-testnet.md): Addresses, blocks, transactions, roles, and constructor inputs of the three Monadeum contracts on Monad testnet (chain id 10143), copied from the committed deployment record.
- [Verify it yourself](https://docs.monadeum.xyz/deployments/verify-yourself.md): Commands with Foundry cast and viem that re-check the Monadeum testnet deployment from the public Monad testnet RPC, without trusting this documentation.
- [Contract interfaces](https://docs.monadeum.xyz/reference/contract-interfaces.md): The Solidity interfaces of AgentSpendPolicy and AgentReallocation as specified in ARCHITECTURE.md, the x402 validator's signature layout, and the rule that money crosses a boundary as an integer.
- [Policy rule schema](https://docs.monadeum.xyz/reference/policy-rule-schema.md): The signed PolicyDocument that parameterizes the risk engine, the proposal and plan types that cross from strategy to execution, the violations it reports, and how the schema versions.
- [Genome hash](https://docs.monadeum.xyz/reference/genome-hash.md): How Monadeum canonicalizes and hashes a genome, a strategy version, a policy, or an execution plan: RFC 8785 JSON canonicalization, a domain tag, keccak256, strict normalization rules, and a fixed test vector.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.