It's a tradeoff. Technically, you need very few programs, you can let an agent do everything and coordinate everything. But that is also inefficient, it's slow and uses a lot of tokens. So you allow the agent to build tools and coordinate those tools, just like we humans do. However, with agents, the threshold of pain is much higher, we can let agents do thing's "manually" where humans would build automations much sooner.

Exactly. Tools like grep, ls are also in the same category. Even in algorithm, we have techniques like memoization and Dynamic Programming which allow us to speed up things. Why should LLM fill up its context by "manually" doing what wc or ls does for you deterministically ?