WeChat Video Account Quick Analysis

The Problem It Solves

In WeChat Video account operations, quickly evaluating an account's content strategy, audience characteristics, and performance is a common need. Traditional analysis often relies on manual data collection or switching between multiple tools, which is time-consuming and lacks comprehensiveness. This skill targets this specific scenario with a lightweight analysis solution: input an account name, and within minutes, generate a structured Markdown report to help operators or analysts determine whether the account is worth benchmarking or learning from, thereby supporting data-driven decisions.

How the Skill Works

This skill is based on an asynchronous task API, with core capabilities and key steps as follows:

• Authentication and Preparation: Read LY_API_KEY from the config.json in the skill directory, or guide the user to obtain it. The API Key is used to send requests to the backend service, ensuring secure access.
• Information Collection and Verification: The user must provide a WeChat Video account name (nickname). The skill first repeats it back for confirmation to avoid errors due to typos or ambiguity. The account name is the sole basis for positioning the analysis; links or video files are not supported.
• Points Deduction Confirmation: Before execution, clearly inform the user of the estimated points deduction (about 120 points) and proceed only after user confirmation. The deduction is based on actual task complexity, with final settlement by the server.
• Asynchronous Task Creation and Polling: Create a task via POST /api/v1/common-gateway/analysis-skill/wx-video-account-quick-analysis, which immediately returns a task_id. Then use the GET endpoint to poll progress, updating the user on status after each poll (e.g., in progress, completed). This process avoids session interruption, ensuring reliable execution of asynchronous tasks.
• Report Delivery: Upon successful completion, extract the pre-rendered report content from data.markdown and render it directly to the user. The report includes modules such as Account Overview, Audience Profile, Content Matrix, Viral Formula, Benchmark Competitors, and Operational Recommendations, presented in Markdown format for easy reading and sharing.
• Error Handling: Take corresponding measures based on the exit code table (e.g., exit code 12 indicates task failure). On failure, inform the user that points have been refunded and ask if they want to re-create the task with the original account name (re-billing), as this API has no retry endpoint.

Applicable Boundaries and Considerations

• Account Name Accuracy: Must use the exact WeChat Video nickname; it is recommended to copy it from WeChat. A single character mistake could lead to analyzing the wrong account or failure.
• Asynchronous Nature: Tasks are executed asynchronously and require multiple polls (every few seconds) to retrieve results. Avoid synchronous blocking calls, ensuring the assistant actively continues polling.
• Transparent Points Deduction: Points are only settled on task success; they are refunded on failure. Users must explicitly confirm before execution, without skipping this step.
• Dependency on Remote Service: Analysis is based on server-side simulation, and results are for reference only, not promising precision to every data point. Major operational decisions are recommended to be cross-checked with the account's actual data.
• Interface Independence: The task_id for this skill is not interchangeable with the full account analysis API; do not mix them.