適合 AI 智能體的軟件堆疊

Most software has two layers. The app and its business logic at the bottom. The agent at the top. Nothing in between.

So the agent reaches straight into the interface. It reads pixels meant for human eyes. It clicks buttons meant for human hands. Every permission decision is invisible, because there was never a layer whose job was to make it.

An agent-ready stack has three layers. The app at the bottom. The agent at the top. And an access layer in the middle. The part that decides what the agent can see and touch.

That middle layer is where the real work lives. Structured reads with field-level scoping, so one record and every record are different doors. Writes that wait for a named human before anything leaves. Every step carrying its permission decision, visible in the log, auditable after the fact.

This is the stack the last two weeks of posts have been building toward. The translation layer. Permissions and approvals. Structured access instead of scraping. Job descriptions instead of tools. The middle layer is where all four live.

Where would you draw the middle layer in your stack, and what would it refuse first?

AIAgents #AgenticAI #DevCommunity