CLAUDE CODE AGENTS: ADVANCED ORCHESTRATION - Chapter 6, Exercise 1
Why a Planning Agent Is Denied Edit Access
====================================================================================

QUESTION:
Explain why a planning agent is denied Edit access, using the same
underlying reasoning this chapter connects to code review.

SOLUTION / EXPLANATION:

Fundamentals' Chapter 6 established why a code-review agent is denied Edit
access: keeping the act of finding a problem separate from the act of
fixing it, so the review stays a complete, honest, and reviewable account
of what's wrong, rather than issues getting silently patched instead of
reported.

This chapter applies the exact same underlying logic to planning, one step
earlier in the process: a planning agent's whole value depends on the plan
existing as a discrete, reviewable proposal BEFORE any implementation
happens - the entire point of separating planning from implementation
(as this chapter explains) is that a flawed approach can be caught and
revised while it's still just a written plan, cheap to change, rather than
after real code has already been built around it.

If a planning agent had Edit access, it could start quietly implementing
alongside - or instead of - actually proposing and writing out a plan,
collapsing the two deliberately separate steps back into one. This would
undermine the exact benefit this chapter says planning provides: there
would no longer be a genuine, discrete "plan" to review and approve before
implementation begins, since the agent could have already started building
something while nominally "just planning."

So the reasoning is structurally identical to Chapter 6's own reviewer
case, just applied to a different pair of steps: keeping "propose an
approach" cleanly separate from "build it," by making sure the proposing
agent structurally cannot also build, guarantees that a genuine review-
before-implementation step actually exists rather than being silently
skipped.

--------------------------------------------------------------------------
WHY THIS WORKS AS AN ANSWER:
It explicitly draws the parallel to Chapter 6's own reviewer reasoning
(separating finding from fixing) and explains precisely how the same
separation principle applies to planning vs. implementing, rather than
treating the restriction as an unrelated new rule.