The Orchestrator

By now Maya’s stack should sound familiar from Chapters 1, 4, and 5 — instructions, project rules, skills, references, tools, plugins, memory, dynamic context, subagent briefs, hooks and CI, evaluation. What is new here is the diagnostic move.

When the system produces a bad grouping, she does not start by rewriting the prompt. She works backward through the run, asking which conversations entered, which skills and memory were active, what the handoff carried, which tool ran, whether security review fired, and whether the eval set contained a case anything like this one. The answers map to layers:

• Evidence missing → context or retrieval.
• Evidence present, procedure wrong → skill.
• Wrong service called → tool description or capability boundary.
• Right recommendation, wrong branch touched → permission.
• Green CI on a test that never represented the customer → evaluation.
• Bad assumption passed downstream → handoff.

The single-platform version of that diagnostic view is shipping; the version that matters for this book is still missing. Either someone assembles the building blocks Chapter 16 describes into the unified view, or a platform vendor builds the whole stack end to end — the way Oracle, Anthropic, or OpenAI would build it, and then owns the entire loop. Which path wins is an open question. That it will be needed is not: the ability to diagnose multi-agent failures, to oversee interactions, and to see the decision chain with its instructions, data, and context at any point in time is the gate between orchestration as a productivity story and orchestration in mission-critical areas — power grids, hospital systems, banks.

The better analogy for all of this is a CI pipeline and a tighter OAuth scope, not a sterner paragraph. You teach a system that can invent its own steps by changing the path it is required to walk, and then by deciding whether a proposed lesson — a memory, a revised skill, a new test, a narrower permission — is safe to keep.

So the question to carry forward is not how to complete one task with an assistant attached. It is what environment would let a bounded system perform this task repeatedly: context, skills, tools, state, evaluation, authority. A prompt helps one run, and the rest of the layers carry classes of runs. That is why this chapter sits ahead of delegation contracts and topologies. You need to know what is actually being designed before anyone can argue about how to design it.

HQ 6 — Assembled. The human and AI each wrote portions of this chapter. I assembled, reviewed, and take responsibility for the whole; the voice and arguments are mine, and I know which parts are which.