CLAUDE CODE AGENTS: ADVANCED ORCHESTRATION - Chapter 2, Exercise 3
Why Converting Every Agent to a Programmatic Definition Is Questionable
====================================================================================

QUESTION:
A developer converts every agent in their project to a programmatically-
defined SDK agent, even ones whose configuration never changes between
runs. Using this chapter's own warning box, explain what's questionable
about this decision.

SOLUTION / EXPLANATION:

This chapter's warning box is explicit that programmatic definitions add
real complexity - more code to write and maintain, and more surface area
for something to go subtly wrong in how the configuration actually gets
assembled - and that this added complexity is worth taking on specifically
when the dynamic behavior it enables is genuinely needed, not as a general
upgrade applied by default.

For an agent whose tools, model, and system prompt never actually change
between runs, none of the genuine benefits of a programmatic definition
apply at all - there's no conditional logic that needs to run, no runtime
variable that needs filling into the prompt, nothing that a static file
couldn't already represent perfectly well. Converting such an agent to a
programmatic definition anyway means taking on the added complexity this
chapter warns about - more code, more potential for a subtle bug in how the
configuration is built - while gaining absolutely nothing from it in
return, since the resulting configuration is identical every single time
regardless of whether it's read from a static file or assembled by running
code.

This chapter's own guidance is to default to hand-written definitions
unless dynamic behavior is genuinely needed, precisely to avoid taking on
this kind of unnecessary complexity. Converting every agent as a blanket
policy - rather than case by case, based on whether each specific agent
actually needs runtime variation - is exactly the pattern this chapter
advises against.

--------------------------------------------------------------------------
WHY THIS WORKS AS AN ANSWER:
It explains specifically why a fixed-configuration agent gains nothing from
a programmatic definition (no genuine dynamic behavior exists to justify
it), while still taking on the real complexity cost the warning box names,
and ties this back to the chapter's own explicit "default to hand-written
unless needed" guidance.