THE ANATOMY OF A PRODUCTION HARNESS (COMPONENT MAP)

ARTIFICIAL INTELLIGENCE ASSISTANT PROMPTS [START OF PROMPT] Produce a component map, as a text-based diagram of boxes and arrows, of this harness, showing the instruction manager, context builder, model adapter, tool registry, permission resolver, budget tracker, and logger, with arrows for the order of operations in one loop iteration: [PASTE LAB REPO FILE LIST OR ARCHITECTURE NOTES HERE]. Flag any missing component from this standard set and any component that appears to be missing error handling. Output format: the diagram followed by a bulleted gap list. [END OF PROMPT] [START OF PROMPT] For this new agent task: [TASK DESCRIPTION, for example "triage incoming support emails"], tell me which of the eight core harness components (instruction manager, context builder, model adapter, tool registry, permission resolver, budget tracker, logger, reporter) I actually need for a first production-safe version, in what order they run in one loop iteration, and which I can safely defer to a later version. Output format: a build-order checklist with a one-line justification per component, and a separate deferred list. [END OF PROMPT]