How LockStock Works

A cryptographic audit chain for AI agents — real-time verification, zero-TTL credentials, encrypted transcripts

What LockStock Does

LockStock maintains a cryptographic hash chain for every AI agent you deploy. Every action the agent takes — a chat completion, a tool call, a file write — extends the chain by one position. This gives you three things no other system provides simultaneously:

The Agent Stays Oblivious

This is the most important design decision in the protocol: the agent does not know LockStock exists.

The agent sends a normal API request to the gateway. The gateway stamps the request (advancing the chain), forwards it to the upstream provider, and returns the response. The agent never handles a credential, never calls an auth endpoint, never manages a token lifecycle. The gateway handles everything.

This means adoption does not require modifying your agents. No SDK. No code changes. No credential management logic. Point your agent at the gateway instead of directly at the provider. The audit trail, the verification, and the encryption happen transparently.

How Verification Works

Every stamp produces a state_hash — a SHA-256 hash that depends on the agent's entire action history. This hash is the credential. It has never existed before, it will never be valid again, and it was not issued by anyone. It is a mathematical consequence of the chain advancing.

The Three-Step Exchange

Step 1 — Stamp. The gateway stamps an action. The server advances the chain and returns the new state_hash.

Step 2 — Present. The gateway attaches the credential to the outgoing request: Authorization: D3cipher {agent_id}:{state_hash}. The agent is oblivious.

Step 3 — Verify. The receiving service makes one GET call:

GET https://api.d3cipher.ai/v1/verify/{agent_id}?hash={state_hash}

The response is one word:

Status	Meaning	Action
current	Credential matches the chain head	Accept
stale	Agent has advanced past this hash	Reject
frozen	Operator hit the kill switch	Reject
revoked	Agent permanently decommissioned	Reject
unknown	Agent not registered	Reject

if status == "current" — that is the entire integration. Everything else is a rejection. The secure behavior is the default behavior.

What Dies When the Credential Dies

The instant the agent takes its next action, the previous credential becomes stale. There is no TTL, no expiry window, no grace period. The credential is valid for exactly one chain position.

Mechanism	Valid For	Compromise Window
API Key	Until manually rotated	Hours to days
OAuth Token	1 hour (typical)	Up to 1 hour
JWT	Minutes to hours	Full TTL window
Atomic Credential	Until next action	Zero

Freeze an agent on the dashboard. The next verification call — from any counterparty, anywhere — returns frozen. Propagation delay: zero.

Nine Trust Relationships, One Protocol

LockStock is not a point solution for one auth pattern. The same protocol — one hash in, one bit out — covers every trust relationship in an AI agent deployment.

Encrypted Transcripts

Every prompt and response passing through the gateway is encrypted with AES-256-GCM envelope encryption before leaving your network. The d3cipher server stores only ciphertext. Your Account Key — generated in your browser, never sent to the server — is the only way to decrypt.

This is not "encryption at rest" on someone else's server. The server literally cannot read your data. Even if the server is compromised, the attacker gets encrypted blobs and wrapped keys — neither of which is useful without the Account Key that you hold.

To audit: open the Auditor tab in the dashboard, enter your Account Key, and read the full content of every conversation — decrypted entirely in your browser.

Five Verification Layers

Authentication is one layer of a five-layer verification stack. Each layer answers a different question. They share the chain but not each other's inputs.

The atomic credential emerges from Layer 1. The state_hash is the chain head after each stamp. Layers 2–5 provide additional guarantees — but the credential itself requires only Layer 1. Each layer is independently verifiable.

Compliance

Because the credential and the audit record are the same atomic operation, the authentication system is the compliance system. There is no separate logging pipeline to keep in sync.

Framework	Requirement	How LockStock Satisfies It
SOC 2 CC7.2	Monitor system components for anomalies	Every agent action is a chain entry. Anomalies break hash linkage. Verified by 6-pass verification kernel.
HIPAA §164.312(b)	Audit controls for information systems	Immutable audit trail. Envelope encryption ensures the server cannot read PHI.
PCI-DSS 10.5	Secure audit trails against alteration	SHA-256 chain — altering any row invalidates all subsequent hashes.

Getting Started

• Gateway Quickstart — Deploy a gateway and see your first audit trail
• CLI Reference — Provision agents, manage keys, audit chains
• API Reference — Endpoints, authentication, verification tiers
• Compliance Quickstart — Independent hash verification for auditors
• Verify Playground — Try the verification protocol live
• Whitepaper — Full protocol specification