KYA Documentation

Start with what you need.
Skip what you don't.

Install, integrate, deploy, understand. Four entry points, in roughly the order you'll need them.

At Runtime ↓

Know Your Agents.
Know Your Principal.

Trust, policy, and signed evidence for every AI agent action.

~ — bash

→ 5 MIN

Quick start

Install. Score one agent. See the breakdown. No DB, no setup. Pure-function entry point.

Open quick start →

→ 20 MIN

Adapter guide

Drop into LangChain, OpenAI Agents, Claude Agent SDK, or any framework. Native callback hooks. No code changes.

Pick your framework →

→ 1 HR

Production deploy

PostgreSQL + Sigstore + signed-recommendation flow. Trust anchor configuration. Operator approval workflow.

Production guide →

→ DEEP

The paper

Threat model, formal lemma (only-tighten algebra), measured benchmarks across 4 backends, 1,200 adversarial probes, the Liang topology-attack analysis.

Read the paper →

Start with what you need.Skip what you don't.