WeChat Video Account Diagnosis and Analysis (Paid Version)

The Problem: Challenges in Analyzing WeChat Video Accounts

Manually analyzing data for WeChat Video accounts is a cumbersome process. Operators or researchers often need to repeatedly check account homepages, tally videos individually, and rely on experience to assess content strategies. This approach is time-consuming and risks missing key data points, such as audience profile details, content matrix distributions, or patterns for viral formulas. For teams needing to benchmark competitors or formulate operational plans, the lack of structured reports can slow down decision-making.

How the Skill Works: An Automated Flow from Input to Report

This skill automates the analysis process by invoking a remote paid decomposition service. Its core capability lies in using the script scripts/analyze_account.py to handle the HTTP call chain, including task initiation, status polling, and report delivery. Key steps are as follows:

• Input and Authentication: The user provides the account name (nickname) of a WeChat Video account. The skill first checks for the LY_API_KEY in the local config.json or environment variable. If missing, it guides the user to obtain it from the Lingyi Data platform and ensures secure storage.
• Estimation and Confirmation: Before full decomposition, the skill runs --estimate-only to create a task, obtaining a task_id and estimated point deduction (typically around 168 points). It reminds the user about account name accuracy and clearly states the point cost, waiting for user confirmation before proceeding to avoid errors.
• Task Execution and Progress Feedback: After confirmation, the skill polls the server using the task_id. During this, it continuously relays progress lines from stderr (such as status, elapsed time) in plain language (e.g., "Decoding video, waiting for about 2 minutes"), ensuring the user is informed throughout. Polling intervals are configurable, defaulting to every 8 seconds.
• Report Generation and Delivery: Upon task completion, the skill extracts the Markdown content between === ACCOUNT_REPORT_START === and === ACCOUNT_REPORT_END === from stdout and renders it directly in the conversation. The report covers dimensions like account overview, audience profile, content matrix analysis, video breakdown, viral formula, competitive benchmarking, and operational suggestions. It also informs the user of the actual point deduction and the report file path.

The entire process relies heavily on the remote service, with the script handling complex network interactions, error retries, and timeout recovery. Users only need to provide an account name to receive a structured analysis report.

Scope and Considerations

This skill is not a universal tool; its boundaries should be understood when using it:

• Input Limitations: It only accepts a WeChat Video account name (nickname) as input, not direct processing of links or video files. The account name must be accurate; misspellings may lead to decomposition failures or wrong accounts.
• Cost Awareness: Each new task incurs point deductions, and the per-task consumption is significant (estimated around 168 points, with possible fluctuations). The skill enforces a second confirmation to clarify costs, and users must ensure sufficient balance in their 01Claw account.
• Network and Stability: The skill depends on a remote service; network fluctuations or server throttling may cause task failures (exit codes 5 or 6). Points are returned in such cases, but manual retry is required. Timed-out tasks can be resumed via --task-id, though not always successfully.
• Report Source: The report is generated as structured JSON by the server, which the script renders into Markdown. Report quality depends on the server's analysis; the skill only delivers it and does not guarantee absolute data accuracy.
• Authentication Security: API keys are stored locally in config.json or environment variables and must be managed properly. If an SSL certificate error occurs, setting LY_SKIP_SSL_VERIFY=1 can be used temporarily, but with awareness of man-in-the-middle risks.

In summary, this skill is suitable for engineers or operators needing quick, deep analysis reports of WeChat Video accounts, but one should fully understand its paid model, input requirements, and network dependencies.