fix(channel): chunk Telegram markdown before rendering

[codeaholicguy]

Summary

• chunk Telegram markdown source with marked tokens before rendering HTML
• render every chunk independently with the existing markdownToTelegramHtml renderer
• split oversized code fences, lists, paragraphs, and emoji-heavy text under Telegram limits
• add lifecycle docs and adapter coverage for long fenced code, nested-list code, paragraphs, emoji length, and unchanged normal markdown

Validation

• npm --workspace @ai-devkit/channel-connector test -- src/tests/adapters/TelegramAdapter.test.ts
• npm --workspace @ai-devkit/channel-connector test
• npm --workspace @ai-devkit/channel-connector run typecheck
• npm --workspace @ai-devkit/channel-connector run lint
• pre-commit hook: repo lint passed with existing warnings; repo tests passed 70 files / 821 tests after building workspace package outputs

Risks

• Markdown chunking now repeatedly renders candidates for validation; acceptable for Telegram-sized sends.
• Plain/source fallback remains for pathological chunks that still cannot fit after semantic splitting.