Read the argument behind the mechanism.
Selected essays by Loren Angoni explore why action-time governance matters. The complete articles remain canonical on langoni.me.
Technical and structural context for 5D.
These are ordinary external links. 5D does not copy or republish the article bodies.
Runtime Is the Moment of Consequence in AI
Why governance becomes real when an AI agent is about to act, and what historical safety failures teach about that boundary.
When AI Governance Reaches the Execution Boundary
How governance intent becomes an action-time decision and a record that can be inspected later.
The Model Can Be Probabilistic. The Authority Boundary Should Not Be.
A technical argument for combining deterministic control, typed judgments, and fail-closed review in high-risk agent paths.
How Do You Test a Control That Can Never Fail?
A practical guide to mutation, path evidence, and an explicit unevaluable result for safety and policy controls.
The Infrastructure Consequential AI Agents Need
A wider view of bounded authority, evidence, and recourse for agents that can affect money, data, rights, or services.
AI Agent Approval Fatigue Is Real
Why asking for approval on every harmless action weakens the human-control loop, and what a policy layer changes.
How Open Source Saved the Internet
How open protocols, shared software, and viable ecosystems kept the internet contestable, and why AI faces a harder version of that test.