fix(memory): account for CJK characters in QMD memory chunking

The QMD memory system uses a fixed 4:1 chars-to-tokens ratio for chunk
sizing, which severely underestimates CJK (Chinese/Japanese/Korean) text
where each character is roughly 1 token. This causes oversized chunks for
CJK users, degrading vector search quality and wasting context window space.

Changes:
- Add shared src/utils/cjk-chars.ts module with CJK-aware character
  counting (estimateStringChars) and token estimation helpers
- Update chunkMarkdown() in src/memory/internal.ts to use weighted
  character lengths for chunk boundary decisions and overlap calculation
- Replace hardcoded estimateTokensFromChars in the context report
  command with the shared utility
- Add 13 unit tests for the CJK estimation module and 5 new tests for
  CJK-aware memory chunking behavior

Backward compatible: pure ASCII/Latin text behavior is unchanged.

Closes #39965
Related: #40216

src/auto-reply/reply/commands-context-report.ts

@@ -5,14 +5,11 @@ import {
 } from "../../agents/pi-embedded-helpers.js";
 import { buildSystemPromptReport } from "../../agents/system-prompt-report.js";
 import type { SessionSystemPromptReport } from "../../config/sessions/types.js";
+import { estimateTokensFromChars } from "../../utils/cjk-chars.js";
 import type { ReplyPayload } from "../types.js";
 import { resolveCommandsSystemPromptBundle } from "./commands-system-prompt.js";
 import type { HandleCommandsParams } from "./commands-types.js";
 
-function estimateTokensFromChars(chars: number): number {
-  return Math.ceil(Math.max(0, chars) / 4);
-}
-
 function formatInt(n: number): string {
   return new Intl.NumberFormat("en-US").format(n);
 }