An agent that can't remember yesterday's correction will repeat the mistake, and one with more access than its job needs can do more harm than its job requires, which is why Fold's privacy reviewer has no shell at all. This thread follows that question through Fold's review panel, where each reviewer's access is part of its design, and this lab, where rules in files stand in for memory.
What it knows and can reach
What should an AI know and be able to reach, and who decides?
context engineering · retrieval · tool use · MCP · memory · permissions · data minimisation
Builds
- EXP-007Building · Pilot
This lab, rebuilt
Nothing goes live unread: building a public site with AI agents, without losing control
This lab, rebuilt in three days by Claude Code from briefs I worked out with Claude. The setup that keeps an agent-built site honest, and the rules that came from mistakes.
- EXP-006Building · Pilot
Fold's review panel
Never mark your own homework: independent AI reviewers for a solo builder
Independent AI reviewers for an app I was building faster than I could read its code. They caught bugs that had passed the tests, and one finding changed what Fold is.