CLAUDE CODE AGENTS: ADVANCED ORCHESTRATION - Chapter 3, Exercise 1
Research, Fix, Changelog: Which Steps Can Run in Parallel?
====================================================================================

QUESTION:
A workflow has three steps: research a bug, write a fix, then write a
changelog entry describing the fix. Explain which of these steps can run
in parallel and which must run sequentially, using this chapter's own
deciding question.

SOLUTION / EXPLANATION:

This chapter's deciding question is: does the second step need the first
step's actual result before it can start? Applying that question to each
adjacent pair in this three-step workflow:

RESEARCH -> WRITE THE FIX:
Writing a fix genuinely requires knowing what's actually causing the bug -
that's the entire point of doing the research first. The fix-writing step
cannot meaningfully begin until the research step's actual findings are
available; it isn't guessing at a plausible cause, it needs the real
answer. This pair must run SEQUENTIALLY.

WRITE THE FIX -> WRITE THE CHANGELOG ENTRY:
A changelog entry describing the fix needs to know what the fix actually
was - what changed, and what user-visible effect it has. This cannot be
written before the fix itself exists, since there would be nothing real yet
to describe. This pair must also run SEQUENTIALLY.

So in this specific three-step workflow, there is no genuine opportunity
for parallel execution at all - each step's job is to use the real, actual
output of the step immediately before it, not a guess or a placeholder.
The entire chain (research, then fix, then changelog) must run
sequentially, one step fully completing before the next begins.

This is a useful illustration that not every workflow with multiple agents
has room for parallelism - sometimes, as here, every step in the chain has
a genuine dependency on the one before it, and sequential execution isn't
a missed optimization, it's the only structure the actual dependencies
allow.

--------------------------------------------------------------------------
WHY THIS WORKS AS AN ANSWER:
It applies the chapter's own deciding question explicitly to each adjacent
pair of steps rather than assuming an answer, and correctly concludes that
this particular workflow has no genuine parallelism opportunity at all,
since every step is a real, direct consumer of the one immediately before
it.