[Discussion] Split core.py into focused submodules?

Question

adpentest/core.py is now ~11,200 lines. Should we split it into submodules, or is single-file the intentional design?

Current single-file rationale (per CLAUDE.md)

• No circular imports
• Centralized tool registry, error handling, profiler
• One file to modify for tool changes
• Easy grep/read for LLM tools

Cost as of today

• Slow IDE navigation, jump-to-def latency
• Merge conflicts concentrate on one file
• Cognitive load for new contributors
• Hard to unit-test in isolation

Options to weigh

1. Keep single-file — status quo, invest in section banners + # region markers instead
2. Light split — extract long-standing self-contained blocks (email protocols, SPN enum, DC detection) into 3-4 submodules, keep core.py as the coordinator
3. Full package split — dc/, enum/, email/, kerberos/, tools/, parallel/ — matches the CLAUDE.md architecture headings

Note: the adpentest_onefile.py single-file distribution can be regenerated automatically (see #40), so a package split does not lose the single-file consumption path.

Non-goals

• Behaviour change
• Public API rename

Cast opinions in the thread.

Start by reading adpentest/core.py and the current rationale in CLAUDE.md, then review issue #40 for the single-file regeneration path. Compare the keep, light-split, and full-package options against the stated costs and non-goals. Done means the thread reaches and records a clear architectural decision.