Agent Execution Proof: a structured audit record for every AI tool call

As AI agents move from demos into production, a quiet assumption breaks down: that logs are sufficient for accountability. They're not. This post explains why, and describes the Agent Execution Proof