Image-Based Speech Writing: Luo Zhenyu Style

Problem

When the input is only an image, quote, or vague topic, it is hard to turn it into a structured long-form commentary with verified background, cases, and a speaking rhythm. This skill targets that gap: not simple rewriting, but expanding sparse input into a 10-minute-plus deep-dive script for readers who expect logical analysis.

How It Works

The workflow is oriented toward Luo Zhenyu-style commentary and produces a speech-like article through several controlled steps:

• Understand the input: identify the core topic, audience, and angle, with WeChat public account readers as the default target.
• Research supporting material: collect news context, industry data, historical examples, and quotes; citations must be verifiable, and fabrication is not allowed.
• Plan the structure: organize a 10-minute-plus delivery into opening hook, problem analysis, case support, deep insight, and closing elevation.
• Apply style constraints: use first-person narration, colloquial rhythm, analogies, rhetorical questions, and measured humor while keeping the content readable.
• Output the script: generate a title, natural paragraphs, [pause] markers, and source citations, optionally saving a UTF-8 file named LogicTalk_[Topic]_[Date].md.

Boundaries

It is best suited for deep-explanatory content rather than brief news updates, formal memos, or marketing copy. If the input is too vague, the skill still needs to fill in research and structure first. The save path in the provided description is macOS-specific, so environments with different desktop locations may need adjustment.