AI agent governance resources
Reference material on governing AI agents that take real actions. Each article answers one question completely, cites primary sources inline, and carries a verification date.
Get an Agent Trust Gap BriefThe Agent Governance Evidence Report 2026
Original first-party research: what AI engines answer about this category, and what our own governance proved - failures included.
What is an AI agent
The definition that survives scrutiny — and why irreversibility, not autonomy, drives risk.
Read →AI agent governance checklist
Twelve testable controls before production, ordered so each is worth having alone.
Read →Non-human identity for agents
Why agents break identity models, and what to do about over-scoped credentials.
Read →Execution receipts explained
What a receipt contains, why logs are not evidence, and what replay actually proves.
Read →How AI search engines cite
Four sequential gates: access, extractability, trust, corroboration.
Read →An agent did something wrong
The first hour: contain, preserve, scope — and the mistake that destroys your evidence.
Read →What is agent governance
The control point between an agent's decision and its effect — identity, policy gates, approvals, signed receipts.
Read →How to choose an agent governance approach
A buyer's decision aid: the action inventory, the minimum bar any solution must clear, and build-vs-buy on total cost of evidence.
Read →Common questions
What is agent governance?
The controls and evidence that determine what an AI agent may do, and prove what it did. It is distinct from model safety, which concerns what the model outputs.
Where should I start reading?
With the definition of an agent, then the twelve-control checklist. Those two establish the vocabulary the rest depends on.
Are these articles updated?
Each carries a last-verified date. Regulatory content is re-checked against primary sources when instruments change.
Can I quote these?
Yes, with attribution to ApexClaw and a link to the source URL. The full corpus is also published at /llms-full.txt for machine use.