#### describe

In this era of scarce attention, falling behind means missing out on traffic. Leveraging AI to scan for **weak signals** that are taking off but haven't yet been noticed by the public within the past 72 hours, and analyzing their growth rates, we predict the next potential viral topic and generate a countdown-style action list, transforming you from a "follower" into a "trendsetter."

#### Core Task

For the user-specified target area $material (e.g., Claude Skill). Objective: Scan fringe data on YouTube, Medium, Google News, and Reddit to identify the three topic directions that have emerged in the past three days with "viral potential," generate three high-click-rate headlines, create a "72-hour sprint to-do list," and compile all the research and conclusions into a document.

Before starting, confirm the target area with the user.

#### Execution Steps

**Step 1: 72-Hour Freshness Scan**

- **Target**: Find outliers that have just emerged in the last 3 days.

- **Action**:

- **Forced Time Window**: When starting a search, the time range must be strictly locked to either "Past 24 hours" or "Past 3 days".

- **YouTube Potential Discovery**: Search for relevant keywords and sort by "upload time". Focus on finding videos from smaller channels (avoiding big influencers) that were posted less than 3 days ago but have already surpassed 2000 views and are generating active discussion in the comments section.

- **Open Network Arbitrage**: Use Google searches for `site:medium.com` or `site:reddit.com` with keywords and "new" to find newly published in-depth discussion threads.

**Step 2: Viral DNA Analysis**

- **Objective:** Analyze why this content went viral. Was it due to emotion? Novelty? Or practicality?

- **Action**:

- **Emotion Extraction**: Analyze the emotional keywords in highly-rated comments (Is it `Anxiety`? `FOMO`? or `Relief`?).

- **Structural Deconstruction**: Summarize the narrative structure of these viral content pieces (e.g., whether they all used a contrast structure of "I gave up X and used Y instead").

**Step 3: Generate a Topic Prediction Report**

- **Goal**: To produce a "topic selection betting list" that not only contains inspiration but also differentiated strategies.

- **Output Content**:

- **Predicting trending topics:** Use `[topic name]` (e.g., Cursor code editor in practice)

- **Signal Source:** For example, "In the past 48 hours, five highly popular videos about 'Cursor's new features' have emerged on YouTube."

- **Emotional hooks**: For example, "a programmer's anxiety about unemployment + a sense of control over new tools".

- **Recommended Title**:

1. *"I deleted VS Code today. Here is why."*

2. *"Why 90% of developers are sleeping on this AI tool?"*

**Step 4: 72-Hour Sprint Checklist (Action Plan)**

- **Objective:** Transform forecasts into urgent execution instructions, with hourly countdowns.

- **Output Format**: **Checklist** For example:

- **🚨 T+0 (Now): Placeholder and Materials**

- Immediately post a short opinion on Knowledge Planet/WeChat Moments to test the waters (attach a draft of the reference text).

- Register/download the relevant tools and take screenshots of 3 key interfaces that seem counterintuitive.

- **🛠️ T+24 (Tomorrow): Production and Differentiation**

- Record/write the main text. Differentiation strategy: Don't create instruction manuals, create "crash tests".

- Create Thumbnail: Use a high-contrast color scheme, referencing the style `[link]`.

- **🚀 T+72 (Launch): Release and Distribution**

- Video/article publishing.

- Pin the "One-Sentence Summary" in the comments section and encourage subscription.

- Go to the original posts found in Step 1 and interact to attract traffic: "I also tested this and found a big bug..."