docs(CLAUDE.md): inline communication principles; drop .junie/ pointer

Two reasons:

1. Survives a laptop switch. The principles previously lived only in
   .junie/guidelines.md; that file is per-machine tool config.
   Centralising in CLAUDE.md (which IS tracked) means the rules
   travel with the repo instead of with the workstation.
2. Single source of truth. Other AI assistants pointed at this repo
   should defer to CLAUDE.md, not maintain their own copies that drift.

The .junie/ dir becomes a per-machine breadcrumb that points back at
CLAUDE.md, and is .gitignore'd in a separate commit.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

CLAUDE.md

@@ -179,8 +179,15 @@ with, endorsed by, or authorised by Bose.
 
 ## Communication style
 
-When working with a human user in this repo, follow the brief direct-
+When working with a human user in this repo:
-answer norms in [.junie/guidelines.md](.junie/guidelines.md):
+
-prioritise direct answers, avoid reverting to earlier task context
+- **Prioritise direct answers** to the question being asked, even when
-when the user asks something out-of-scope, no assumptions in place of
+  it sits outside the current task or project context. Don't divert
-real information.
+  back to whatever you were doing when the user asks something else.
+- **Don't substitute assumptions for real information.** When something
+  is unclear, ask or check, rather than guessing and proceeding.
+
+These principles also apply to other AI assistants pointed at this
+repo. Tool-specific config dirs (e.g. `.junie/`, `.claude/`) should
+defer to this file as the source of truth instead of carrying their
+own copies.