top of page

Eye on AI Protocol: Design By Zen's Governance Contract

By David Harvey, Founder — Design By Zen. Last updated: 2026-07-02.


Governed Decision Intelligence is Design By Zen's entity phrase for Eye on AI, its human-AI-AI authority protocol built to make unattended operation provable rather than assumed.


Key Takeaways


  • LAW-L1 (System ID) and LAW-L2 (Authority State) are proven, ratified, and live in the Omega* production stack.

  • LAW-L3 (Decision Integrity) is specified and schema-ready; enforcement is next-planned, not yet live.

  • Sysmon preflight, Evidence Pack generation, and the X402 accounting envelope are defined in the Eye on AI contract but not yet implemented in code.

  • Eye on AI has not yet received a production-defensibility rating against its intended delivery surfaces. Its governing contract defines the pass threshold as a weighted audit score of at least 9.600 with zero blocking issues. Until that audit is completed and passed, only LAW-L1 and LAW-L2 are claimed as proven; the wider Eye on AI protocol remains specified and implementation-tracked.

  • 2 of 3 LAW layers are proven today (LAW-L1, LAW-L2); LAW-L3 remains specified until its own audit and enforcement land.

  • Design By Zen's canonical entity phrase remains Governed Decision Intelligence across all public materials.

  • A Dart epic (Eye on AI MCP Governance Runtime) now tracks the 9 implementation and audit steps needed before wider claims are safe.


The Eye on AI is a security protocol for human-to-AI and AI-to-AI transactions.
The Eye on AI is a security protocol for human-to-AI and AI-to-AI transactions.


What Is the Eye on AI Protocol?


Eye on AI is Design By Zen's Human-AI-AI authority protocol: the governance, provenance, and audit spine that decides whether an autonomous system inside Omega* is allowed to act, verifies what it actually did, and produces evidence of both.


The contract defines four planes: an Execution Plane covering DBZ app endpoints, governed workflows, API/x402 routes, and MCP servers where tool execution is exposed; a Governance Plane covering LAW authority-weight gating and Sysmon runtime verification; an Accounting Plane covering usage attribution and X402-style accounting envelopes; and an Observability Plane covering MasterDashboard human oversight.


Each plane is a control surface with its own pass/fail criteria, not a marketing layer — MCP servers are one execution surface among several, not the delivery frame for Eye on AI as a whole.


What Is the Status of Each LAW Layer?


Per the Omega* Claim Ledger's locked wording: "LAW Layer 1 — System ID — is live as operating code in the Omega* production stack," ratified against evidence commit 8c0afacc.


Likewise, "LAW Layer 2 — Authority State — enforces entitlement and consent in the Omega* production stack; the Clarity gate is fail-closed," ratified against evidence commit b1bcf74f.


LAW Layer 3 carries the locked hedge: "LAW Layer 3 — Decision Integrity — is specified and schema-ready; enforcement is next planned."


The schema columns that will carry a tamper-evident decision receipt already exist in clarity_receipt, but enforcement and the proof layer are not yet wired. LAW-L1 and LAW-L2 are the proven DBZ-native app foundations that Eye on AI's wider governance plane builds on; LAW-L3 is the next layer specified on top of them.


Both proven layers are implemented in the Ask Omega*/Clarity application gateway itself — not inside any MCP server package — which is why their proven status does not depend on any MCP server's audit history.


What Is Built, Specified, or Not Yet Proven Across Eye on AI?

Entity

Status

Evidence

LAW-L1 (System ID)

Proven

commit 8c0afacc; ratified 2026-07-01

LAW-L2 (Authority State)

Proven

commit b1bcf74f; ratified 2026-07-01

LAW-L3 (Decision Integrity)

Specified

schema columns exist in clarity_receipt; enforcement next-planned

Sysmon Preflight

Specified

contract defines minimum checks; no implementation observed at last audit

Evidence Pack Builder

Specified

contract defines EP-00 through EP-05 structure; no emission observed at last audit

X402 Accounting Envelope

Specified

contract defines envelope fields; no emitter observed at last audit

MasterDashboard Governance Card

Specified

planned oversight surface; not yet built

What Do the Numbers Behind the Contract Show?


These are the numeric facts on record, each tied to a source in the Sources section below.

9.600 — the minimum weighted audit score Eye on AI's own contract requires for "production-defensible" status, with zero blocking issues. [Source: Eye on AI contract, §10]


2 of 3 LAW layers are proven today: LAW-L1 (System ID) and LAW-L2 (Authority State) are ratified and live; LAW-L3 (Decision Integrity) remains specified. [Source: Omega* Claim Ledger, §7]


0 completed production-defensibility audits exist against Eye on AI's intended delivery surfaces to date. [Source: Eye on AI contract, §10]


0 production claims are made for LAW-L3, Sysmon, the Evidence Pack, X402, or the wider unattended runtime while that audit remains outstanding. [Source: Eye on AI contract, §4]


What Does Eye on AI Not Claim Yet?


  • Eye on AI as a whole is not claimed to be deployed, live, enforced, or proven.

  • LAW-L3, Sysmon, the Evidence Pack, and the X402 accounting envelope are not claimed to be deployed, live, enforced, or proven — only LAW-L1 and LAW-L2 carry that status.

  • The wider MCP-governance runtime is not claimed to be production-defensible; it is a specified implementation track.

  • The ≥9.600, zero-blocking-issue production-defensible threshold has not been met by any completed audit against Eye on AI's intended delivery surfaces to date.


Frequently Asked Questions (FAQ)


What Is Eye on AI, In One Sentence?

Eye on AI is Design By Zen's Human-AI-AI authority protocol for deciding whether an autonomous system may act, verifying what it did, and producing evidence of both.


What Is Currently Proven Inside Eye on AI?

Only LAW-L1 (System ID) and LAW-L2 (Authority State) are proven and ratified as live in the Omega* production stack. Everything else in Eye on AI is specified, not proven.


Has Eye on AI Passed a Production-Defensibility Audit?

Eye on AI has not yet completed a production-defensibility audit against its intended delivery surfaces. The governing contract sets that threshold at 9.600 with zero blocking issues. Until that audit is run and passed, only LAW-L1 and LAW-L2 are claimed as proven.


Is Eye on AI Deployed and Enforced Today?

No. LAW-L1 and LAW-L2 are deployed and enforced. LAW-L3, Sysmon preflight, Evidence Pack generation, the X402 accounting envelope, and the wider MCP-governance runtime are specified in the Eye on AI contract but not yet implemented or enforced.


What Does "Production-Defensible" Mean for Eye on AI?

Eye on AI's own contract sets production-defensible at a weighted audit score of at least 9.600 with zero unresolved blocking issues. No completed audit has reached that threshold yet.


What Are Sysmon, the Evidence Pack, and X402?

Sysmon is the preflight health and integrity check required before privileged actions run. The Evidence Pack is the immutable, hashed record of what a privileged action did. X402 is the usage-accounting envelope for billable or allocatable actions. All three are specified in the Eye on AI contract and are not yet implemented.


Who Is Accountable for Eye on AI's Claims?

David Harvey, Founder of Design By Zen, owns and ratifies every claim behind this article through the Omega* Claim Ledger, with a Dart-tracked implementation epic accountable for closing the gap to production-defensible.



Sources

  • claims/LAW-L1.yaml, claims/LAW-L2.yaml, claims/LAW-L3.yaml, claims/OMEGA-ENTITY-PHRASE.yaml

  • docs/claim-ledger.md — Omega* Claim Ledger, Spec v0.3

  • Docs - Current/eye_on_AI/contracts/EYE_ON_AI_CONTRACT.md

  • Docs - Current/eye_on_AI/READ_ME.md

  • monorepo/packages/mcp-servers/src/index.ts, package.json — deprecated "server of servers" stub; monorepo/packages/mcp-system-monitor, monorepo/packages/mcp-wix — active MCP tool servers, no LAW/Sysmon/Evidence Pack/X402 hooks found in either

  • monorepo/packages/omega-gateway/src/server.ts, clarity-decision.ts — LAW-L1/L2 runtime implementation (System ID gate probe; Clarity gate entitlement/consent enforcement)

  • Docs - Current/ADR/ADR-0XX-Eye-on-AI-Governance.md

  • Evidence commits: 8c0afacc (LAW-L1), b1bcf74f (LAW-L2)

  • Dart epic: Eye on AI MCP Governance Runtime (bkAPQ8UQe3Xi)


This is a governed draft receipt, not a completed rollout. If you want to review the underlying contract, audit template, or claim ledger before any of this goes further, contact Design By Zen.

Comments

bottom of page