#### describe

The debate preparation process is standardized into an industrial-grade assembly line of "divergence-convergence-simulation-output". Starting from understanding the topic and formulating the argument, through AI-assisted logical self-checking and red-blue team competition, a competition-level argument draft that is logically rigorous, data-rich, and has undergone practical simulation is finally generated.

#### Core Task

For the user-inputted "material" (e.g., whether remote work is beneficial to employees' long-term development), as the "material" (for/against), clearly define the "core argument." Objective: Develop 3-5 well-reasoned core arguments, generate a document, search the entire internet for supporting factual data for each argument as evidence, add it to the document, create an interactive webpage to help users conduct "blind box-style attack and defense exercises," and ultimately produce a "practical argument draft."

Before commencing any action, confirm the debate topic with the user. Proceed only after the user responds.

####

**Step 1: Question Analysis and Framework Construction (Structuring)**

- **Objective:** Establish the battlefield boundaries and logical framework.

- **Action**:

- **Positioning Stance**: Clearly define bottom lines and winning criteria.

- **Build the framework:** Create a document listing 3-5 core arguments (results/reasons/value aspects).

- **Logical Self-Check:** Use AI to generate **3 potential logical flaws** (such as "slippery slope fallacy" and "sample bias") for each argument, and write the remedial logic in advance in the document.

**Step 2: Evidence Filling**

- **Goal:** To ground logic in real data.

- **Action**:

- **Agent Sweep**: Launch Agent Research to crawl relevant papers, reports, and news from the entire web.

- **Snip Extraction**: Extract (snip) key data or expert opinions from the source material.

- **Embedded skeleton**: Drag the extracted **Snippet** under the corresponding argument title in the Step 1 document to form a closed-loop structure of "argument + irrefutable evidence + source link".

**Step 3: Blind Box Attack and Defense Simulation**

- **Goal:** Muscle memory training to break free from mental echo chambers.

- **Action**:

- **Build Battlefield**: Generate an offensive and defensive context based on existing arguments and create a webpage.

- **Random Card Draw**: The AI ​​is tasked with playing the role of the "sharp opposing side," randomly selecting one of our arguments for challenging questions. No user input is required; it's simply a card draw, with a 60-second time limit for each answer. The card is automatically redrawn after 60 seconds. Users can also manually redraw the card.

**Step 4: Output the practical argument draft (Scripting)**

- **Objective:** To generate text suitable for on-site reading and quick response.

- **Output**:

- **Visualized Text**: Key quotes are **bold**, and detailed data is hidden using a **Toggle List** to ensure a clean visual experience while remaining readily accessible.

- **Defense Tips**: Attach the "defense golden rules" summarized in Step 3 next to the corresponding arguments.

- **Team-wide synchronization**: Generate a read-only link and share it with teammates to ensure consistency in communication. ››