diff --git a/README.md b/README.md index ace1a89..b5b5fa5 100644 --- a/README.md +++ b/README.md @@ -1,14 +1,12 @@ # 🛡️ secure-agent-patterns -> Defensive patterns for **hardening agentic AI systems** — how to give an LLM tools and autonomy without giving away the keys to the kingdom. +> The **threat model and control-plane design** for governing agentic AI — how to give an LLM tools and autonomy without giving away the keys to the kingdom. This is the intellectual foundation behind [**Vertirite**](https://vertirite.com), SurgeXi's AI-governance control plane. [![ci](https://github.com/tsmith-surgexi/secure-agent-patterns/actions/workflows/ci.yml/badge.svg)](https://github.com/tsmith-surgexi/secure-agent-patterns/actions/workflows/ci.yml) [![license: source-available](https://img.shields.io/badge/license-source--available-blue.svg)](LICENSE) -A practitioner's reference for the security side of agentic AI. As LLMs gain the ability to call tools, browse, and act, the attack surface shifts. This repo collects the **defensive** patterns I apply when designing agent systems — the same posture enterprise AI-security roles are now hiring for. - -> This is defense-only material: it shows how to *protect* agent systems, not how to attack them. +**What this demonstrates.** As LLMs gain the ability to call tools, browse, and act, the attack surface shifts. This repo lays out the security posture I design agent systems around: treat model output as untrusted, route every side-effecting action through a single choke point, deny anything outside its grant, drop egress that isn't on the allowlist, and hold high-risk actions for a human. It's the threat model, the control-plane pattern, and a runnable proof — the *why* and the *how* that Vertirite productizes into a deployable control plane. Defense-only: it shows how to *protect* agent systems, not how to attack them. Setup and the runnable demo are below. ---