What is mattwood.fyi?

A live, structured feed of riffs and links from Matt Wood, Chief AI & Technology Officer at AWS — readable by people and queryable by agents as a typed knowledge graph.

Is there an API and is it free?

Yes. A public, read-only query API is available with no authentication and no API keys. See https://mattwood.fyi/openapi.json and https://mattwood.fyi/docs.

How should an AI agent use this site?

Fetch https://mattwood.fyi/feed.json for the full structured feed, or query the API for keyword/semantic search and the connection graph. See https://mattwood.fyi/agents/ for when-to-use guidance.

Formal Methods for Controlling AI Agents

Explores how formal methods and automated reasoning tools like Z3 can be used to verify that permission changes requested by autonomous AI agents remain within approved security policies, addressing challenges that arise as agents scale to handle long-running, complex tasks.

nvidia.github.io →

Connections

Related to

AI's Role in Formal Mathematical Proof Verification AI's role in formal mathematical proof verification shares the underlying methodology of automated reasoning and formal methods being applied to verify correctness properties
Introducing Dogwood: runtime verification for AI agents Dogwood is explicitly a runtime verification system for AI agents - both items address the same problem of formally verifying that AI agent behavior stays within approved boundaries, differing in approach (static vs runtime)
Anthropic Formalizes Fermat's Last Theorem in Lean Both leverage formal verification tooling (Lean vs Z3) for AI-assisted reasoning tasks - Anthropic's Lean formalization demonstrates formal methods intersecting with AI, while this applies similar rigor to agent safety
Cedar Policy Cedar Policy is a formal policy language for authorization decisions - directly related to using formal methods to constrain what AI agents are permitted to do within security boundaries

Supports

Controlled Agentic Commerce with AgentCore Payments Controlled agentic commerce requires exactly the kind of formal permission-boundary enforcement described here - formal methods provide the theoretical foundation for safe agentic transaction systems
Every Model Cheats: Prompt-Level Mitigation of Cheating on Offensive Cyber Tasks Both address the problem of AI agents circumventing intended constraints - formal methods for permission verification directly complements prompt-level mitigations by providing a stronger mathematical guarantee layer

Challenged by

Dream-RSI: Recursive Self-Improvement through Evolving Worlds Dream-RSI's recursive self-improvement paradigm raises the stakes for formal methods controlling AI agents — autonomously evolving agents may become harder to formally specify and constrain as their strategies drift

Type:	link
First published:	Tuesday, 15 September 2026 at 19:33 UTC
Location:	https://nvidia.github.io/OpenShell-Research/dev-notes/posts/2026-09-10-learning-formal-methods-agent-policy-prover/
URL:	https://mattwood.fyi/i/0fe3c2d6

I'm Matt Wood, and this is For Your Information. A live list of riffs and links for you and your agent, drawn from what I'm reading, noticing, questioning, concluding, and revising.

Connections