Add numbered architecture knowledge docs for OpenKore LLM KB

[boscv]

Motivation

• Provide a concise, structured architecture knowledge layer so an LLM can reason about OpenKore internals, extension points, and high-level runtime flows.
• Focus documentation on critical areas used by automation and analysis: AI, Actor system, Networking, Task system, Plugin/Macro systems, Configuration, and Routing.

Description

• Added six new numbered knowledge files under openkore_llm_knowledge/knowledge/: 01_llm_context.md, 02_architecture_overview.md, 03_system_architecture_map.md, 04_core_subsystems.md, 05_module_dependency_map.md, and 20_code_index.md that together map structure, entry points, and module relationships.
• Each file references concrete modules and directories (for example src/functions.pl, src/AI.pm, src/Actor.pm, src/Network.pm, src/TaskManager.pm, plugins/macro/*, control/*, tables/*, and fields/*) to ground explanations in repository layout.
• Descriptions emphasize runtime flows (main loop -> receive -> actor -> AI -> tasks -> send), coupling hotspots (src/Globals.pm, src/Commands.pm, src/TaskManager.pm, src/Network/Receive.pm) and a recommended navigation order for code review and analysis.

Testing

• Ran an automated path-validation script that extracts backticked module paths from the new docs and verifies referenced concrete paths exist, excluding wildcard/grouped notations; the adjusted validation run succeeded.
• Performed a repository file scan to ensure the new openkore_llm_knowledge/knowledge/ files are present and readable by tooling, and no code or unit tests were modified.

---

Codex Task