examples: short-video narration recipe (silent clip → 9:16 draft with voiceover + captions), en + zh-CN

[renezander030]

What

• examples/short-video-narration.md and examples/short-video-narration.zh-CN.md: silent clip → 9:16 draft with a TTS voiceover and script-accurate captions, as four existing commands (quickstart --ratio 9:16 → tts --text-file → caption --from-segment --script → lint) and as one script.
• examples/scripts/narrate-short.sh chains the four steps: prints the draft path and the voiceover segment, treats a missing whisper as "captions skipped" rather than a failure, stops with the CLI's own error on anything else. CAPCUT_TTS_CMD swaps the engine.
• examples/scripts/edge-tts-wav.sh bridges edge-tts (which only writes MP3) to the WAV that tts expects at {out}, via ffmpeg.
• The vision-model step that writes the narration is optional and stays outside the CLI: the script is a plain text file, one sentence per line.

Why

The most common short-form pipeline in the JianYing ecosystem is footage → narration → captions → app, and every piece of it already existed as a command. This is the recipe that shows them as one path, in the language most of that audience reads first.

Checks

• Script run end to end against the built CLI with a synthetic clip and a stand-in engine: draft created (1080×1920), voiceover segment added, captions skipped cleanly without whisper, lint clean. Name-omitted form, failing engine and missing edge-tts each exit 1 with the CLI's message.
• bash -n on both scripts. No source changes; no test suite impact.

Docs

Examples index row + note that this recipe exists in both languages; a pointer from the zh-CN quickstart's "接下来"; CHANGELOG (Unreleased).