Proof, not promises.
11/11 publishes the exact cryptographic primitives used at every layer of the control plane. Reviewers do not have to guess. Every public claim is reproducible against the live, machine-readable proof endpoint.
The cryptography is named, not implied
Primitives, stated at every layer.
{
"decision": "ALLOW",
"artifact": "ed25519:9b3a…c41",
"sha3_512": "f4c1…7e2",
"blake2b" : "a90d…11b",
"tenant" : "t_8f21…",
"latency_ms": 41
}
--- or ---
{
"decision": "DENY",
"reason" : "policy.boundary.exceeded",
"enforced": true
}
Both paths are evidence
ALLOW and DENY are signed.
A denial is not a silent failure, it is a signed, persisted record with a policy-violation reason. Verify any decision against the live machine-readable proof endpoint.
- ✓ Decision hash and lineage published publicly
- ✓ Reproducible, no demo wrapper, no slide-only diagram
- ✓ Signed typically in under 100ms by the production gateway
EA-11 Evidence State · v2.2 Live
Deterministic, Merkle-rooted execution evidence.
11/11 has upgraded from single proof-hash evidence to deterministic EA-11 Evidence State v2.2. Each governed execution now binds Decision Hash, Artifact Hash, Execution Hash, Audit Hash, and Lineage Root into a Merkle-rooted evidence state with inclusion proof verification and fail-closed tamper detection. This strengthens third-party verification, audit reconstruction, certification readiness, evidence lineage integrity, and governed execution trust.
EA-11 Evidence State v2.2 is live: deterministic component hashes, Merkle-rooted execution evidence, inclusion proof verification, fail-closed tamper detection, real post-quantum signatures, SDK verification, and Ed25519 JWKS discovery.
- ✓ Decision Hash
- ✓ Artifact Hash
- ✓ Execution Hash
- ✓ Audit Hash
- ✓ Lineage Root
- ✓ EA-11 Merkle Root
- ✓ EA-11 Evidence Root
- ✓ Merkle Inclusion Proofs
- ✓ Fail-Closed Evidence Verification
- ✓ Tamper Detection
- ✓ Formal Specification
- ✓ 14 Tests Passing
EA-11 Evidence Verification
Merkle inclusion proof verification implemented; SDK tests passing.
JWKS Origin Endpoint Live
Gateway running with Ed25519 public verification discovery active.
PQ Enforcement Required
Real ML-DSA-65 and SLH-DSA-SHA2-128f enabled; post-quantum verification VERIFIED.
- ✓ EA11_EVIDENCE_STATE_V2_2
- ✓ Real ML-DSA-65 Enabled
- ✓ Real SLH-DSA-SHA2-128f Enabled
- ✓ PQ Verification VERIFIED
- ✓ PQ Enforcement REQUIRED
- ✓ Ledger Integrity 100%
- ✓ Fail-Closed Runtime Enabled
- ✓ PQ Enforcement: REQUIRED
- ✓ PQ Verification: VERIFIED
- ✓ ML-DSA-65: VALID
- ✓ SLH-DSA-SHA2-128f: VALID
- ✓ Evidence State: EA11_EVIDENCE_STATE_V2_2
- ✓ Ledger Integrity: 100%
SDK · In active development
Verify a decision, then govern your own agent.
@11ai/execution-governance is the single developer surface for the 11/11 control plane. One package, one integration, two capabilities. The SDK requests and verifies. The control plane holds the keys and signs. Governance sits on the action, not the model, so one integration covers OpenAI, Anthropic, Grok, and your own custom agents alike. Vendor-neutral by construction.
Check any decision yourself
Verify governance decisions against the public JWKS and proof endpoint. The same public verification surface this site already exposes, now callable from your own code.
Put your agent behind the boundary
Wrap your own agent's actions in pre-execution authorization and fail-closed enforcement, with a signed proof of every decision. The fuller use of the same SDK.
Vendor-neutral by construction
It governs the action, not the model. One integration covers every agent you run.
OpenAI
Anthropic
Grok
Custom agents
How the govern path works
Every action your agent takes is authorized before it can run, and proven after.
Agent builds a request
SDK signs and sends it
Signed allow or deny
Verify against public JWKS
Proceed, or never run
- ✓ Your agent builds an action request.
- ✓ The SDK signs the request and sends it to the control plane.
- ✓ The control plane returns a signed allow or deny decision.
- ✓ The SDK verifies that signature against the public JWKS.
- ✓ Allowed actions proceed. Denied actions never run. Every decision is recorded.
The SDK is in active development. It is not yet generally available, publicly downloadable, or production-supported for outside integrations. That is fine, because the control plane it governs against is live and public today. You do not have to take our word for it, you can verify a real signed decision yourself.
Execution Governance™ · Governed Execution™ · EA-11™ · Execution Arithmetic™ · Execution Evidence State™ · Patent Pending
Published & citable
The doctrine is on the public record.
11/11 publishes its execution governance architecture as formal, DOI-registered reports through the 11/11 AI Research Division. Each is permanently archived, openly licensed, and independently citable. This is proof you can footnote.
11/11 Execution Governance Architecture: A Fail-Closed Execution Governance Layer for Regulated AI and Compute Infrastructure
DOI 10.5281/zenodo.20252295 Read on Zenodo → Report / Architecture11/11 AI: A Deterministic Execution Authority Architecture for Governed AI, Data, and Autonomous Systems
DOI 10.5281/zenodo.20252639 Read on Zenodo → Standard / CharterRFC-EG-0001: Charter of the RFC-EG (Request for Comments, Execution Governance) Series
DOI 10.5281/zenodo.20260882 Read on Zenodo →Reference verifier & standards track
Verify a governed decision against the reference implementation.
Open-source verifier: clone and run in 30 seconds 
RFC-EG-0010 Lineage Verifier
Open on GitHub → Conformance profiles11/11 Governance Profiles
Open on GitHub → Archive11/11 AI Research Division: full corpus
Browse all records →Verification before runtime. Enforcement during runtime. Proof after runtime. Copyright © 2026 11 AI Blockchain Developments Land and IP Trust. Execution Governance™, Governed Execution™, EA-11™, Execution Arithmetic™ , Execution Evidence State™ , Patent Pending.