> For the complete documentation index, see [llms.txt](https://yellowpaper.axblade.io/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://yellowpaper.axblade.io/readme.md). # AxBlade Technical Yellowpaper > **AxBlade** is a Layer-2 ZK Rollup forked from zkSync-era, equipped with a **5-layer AI Agent Governance architecture** (L0 Behavior Recording / L1 Decision Receipt / L2 Policy Engine / L3 Explanation Layer / L4 Regulatory Anchor). Its objective is to commit on-chain — immutably — the behavior, decisions, policy compliance, explainability, and regulatory alignment of AI Agents. *** ## What This Yellowpaper Covers Unlike the marketing-oriented Whitepaper, this Yellowpaper targets **integrators / smart-contract developers / auditors / regulators** and is organized along five dimensions: | Dimension | Section | Audience | | ---------------- | ------------------------------------------------------------------- | ------------------------------------ | | **Algorithms** | [Algorithms](/algorithms/merkle-tree-positional.md) | Protocol engineers / audit firms | | **Use Cases** | [Protocol Mechanics](/protocol-mechanics/pob-behavior-recording.md) | Product / business / legal | | **Architecture** | [Overview](/overview/5-layer-architecture.md) | Technical stakeholders / investors | | **Ownership** | [Registries](/registries-ownership/policy-registry.md) | Compliance / legal / model providers | | **Integration** | [Integration](/integration/sdk-typescript.md) | Agent developers / integrators | *** ## Current Status: PoC LIVE on chain 391 > **Note**: AxBlade is currently at Proof of Concept (PoC) stage and is **not mainnet-ready on Ethereum**. See [Not Mainnet-Ready Statement](/legal-and-disclaimers/not-mainnet-ready.md). | Dimension | Status | | ------------------------- | --------------------------------------------------------------------------------------------------------------------------- | | **Chain** | chain 391 (`https://rpc-testnet.axblade.io`) — anvil-zksync single-node demo | | **Contracts** | 33 production contracts deployed + 1248 tests 100% passing | | **L0 Behavior Recording** | PoB v2 + ChallengeManager + live-chain fraud-proof e2e (0.5 ETH real slash) | | **L1 Decision Receipt** | PoD + ECDSA real signature verification + ZK/TEE mock pending swap | | **L2 Policy Engine** | PolicyRegistry + EU AI Act high-risk template + live-chain SDK e2e | | **L3 Explanation Layer** | ExplanationRegistry + deterministic stub explainer | | **L4 Regulatory Anchor** | EU AI Act policy hash on-chain | | **Audit** | 0 independent audit rounds; self-review prep complete ([audit prep package](/consolidated-resources/audit-prep-package.md)) | | **Stability** | G3 30-min soak: 100% uptime / p50=749ms / 0 drift | | **On-chain facts** | 13 key txs on-chain ([see table](/consolidated-resources/contract-addresses-chain-391.md)) | *** ## Navigation * [5-Layer Architecture Overview](/overview/5-layer-architecture.md) — understand the full picture in one diagram * [Chain 391 Quickstart](/getting-started/chain-391-quickstart.md) — up and running in 5 minutes * [Fraud-Proof Mechanism](/protocol-mechanics/fraud-proof-mechanism.md) — real slash, live chain, e2e walkthrough * [PolicyRegistry Ownership](/registries-ownership/policy-registry.md) — who can register a policy * [TypeScript SDK](/integration/sdk-typescript.md) — viem + @axblade/sdk *** ## Summary 1. **Traditional AI has no tamper-evident behavior ledger** — AxBlade hashes on-chain, immutably, the answers to five questions for every AI Agent: what it did, why it did it, under which policy, whether it can be explained, and whether it is compliant. 2. **Not just another ZK Rollup** — AxBlade layers a **5-tier governance protocol purpose-built for AI Agents** on top of ZK Rollup, aligned with EU AI Act / NIST AI RMF / China Generative AI Service Management Regulations. 3. **PoC is complete; mainnet is 12–18 months away** — on-chain facts are independently verifiable on chain 391; mainnet deployment requires 2–3 independent audit rounds + a real ZK prover + real TEE replacing mocks + legal entity formation. *** ## Document Version * **Yellowpaper version**: v1.0 — 2026-05-09 (commit `7c438db`) * **Corresponding Whitepaper version**: v0.3 (2026-04-30) * **Chain baseline**: chain 391, block ≥ 21055 (2026-05-05 G3 soak) * **Audit prep baseline**: commit `7c438db` (post 6 audit prep commits) *** ## Feedback and Collaboration * Source code: AxBlade repository (to be made public) * Vulnerability disclosure: see [Responsible Disclosure](/security/responsible-disclosure.md) * Audit firm RFQ: see [Audit Prep Package](/consolidated-resources/audit-prep-package.md) --- # Agent Instructions This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com. ## Querying This Documentation If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question. Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter: ``` GET https://yellowpaper.axblade.io/readme.md?ask=&goal= ``` `ask` is the immediate question: it should be specific, self-contained, and written in natural language. `goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal. The response will contain a direct answer to the question and relevant excerpts and sources from the documentation. Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.