Why move away from explicit Three-Role & PTT structure?

[Rexturnull]

I've been analyzing the latest PentestGPT (v1.0+) and noticed a significant shift from the original modular design. I have a few concerns:

1. Monolithic vs. Modular: The core strength of PentestGPT was the explicit separation of Reasoning, Task, and Parsing Handlers. In the current version, this feels merged into a single "black-box" prompt. Is the multi-role logic still explicitly enforced?

2. PTT Visibility: The PTT was a vital reasoning engine for state management. Why was the explicit, structured PTT visualization discarded for an implicit agentic loop?

3. CTF vs. Enterprise: The current prompts seem heavily pivoted toward CTF flag-hunting. This "never give up" mindset is often counterproductive in professional enterprise environments where stability, service availability, and structured risk assessment are prioritized over raw exploitation.

Without the explicit PTT and role structure, how does this differ from using a standard LLM with a long system prompt? I believe transparency in the reasoning process is crucial for professional use.