Research Library

Independent research on AI agents, security, and infrastructure — published by West AI Labs.

Security & Threat Landscape

The Agentic Economy and the Prompt Injection Arms Race

Visa, Mastercard, Stripe, and Google built payment-capable agent infrastructure in 6 months. The injection problem is completely unsolved. Here's why that matters.

The Promptware Kill Chain

Prompt injection is the wrong frame. Attacks on LLM-based systems have evolved into a distinct class of malware — and 90%+ of published defenses have already been bypassed.

The Poisoned Orchestrator Attack: Trust Hierarchy Exploitation in Multi-Agent AI Systems

When the orchestrator is compromised, every sub-agent it spawns inherits that compromise through its system prompt. This is a novel vulnerability class — and no existing monitoring tool can see it.

AI Social Platforms & Agent Behavior

Moltbook and the Rise of AI Social Platforms

The first AI-only social network at scale — and what a massive data breach reveals about building agent infrastructure responsibly.

Infrastructure & Economics

What's Actually Driving Down the Cost of AI Inference

Everyone's talking about better chips. But the real cost reduction in AI inference has been driven by algorithms — and the endgame is local hardware where marginal cost hits zero.

Perspectives

Where AI Actually Lands by End of Year

Not a forecast deck. By Q4, the eval question shifts from 'is it accurate?' to 'does it show up reliably?' Here's what actually changes in 2026 — and what doesn't.

Research maintained as living documents. Last updated: March 11, 2026.