Glossary
Patchrooms glossary
The concepts behind turning visual feedback on AI-built apps into agent-ready patch context.
-
Patch context
Structured feedback a coding agent can act on without guessing.
-
Agent-ready feedback
Feedback structured so a coding agent can act on it without interpretation.
-
AI artifact review
Reviewing an AI-generated preview in place and feeding fixes back to the agent.
-
Loop engineering
Designing the act-observe-decide-repeat cycle an agent runs, not just its prompt.
-
Human in the loop
A person approves one specific step in an agent workflow, at a checkpoint the system defines, not continuous supervision.
-
Review bottleneck
The point where human review, not agent output, limits how fast a loop ships.
-
Agentic loop
The observe-decide-act-verify cycle an AI agent runs on its own, triggered by a heartbeat, cron, hook, or unmet goal.
-
AI slop
Low-quality AI output shipped without review: the default result when agent loops run faster than humans can check them.