SuperAPI AI Directory

text-gen-skill

General & other

View repository

Updated 10-16 08:44

About this skill

AI text generation

Trigger scenarios

Examples of user wording:

  • Article writing: "Help me write a WeChat official account article about coffee culture", "Write a Xiaohongshu note recommending spring drinks"
  • Copywriting generation: "Help me write a slogan for a coffee shop", "Generate ad copy for a new product launch"
  • Code generation: "Write a bubble sort in Python", "Help me write a cloud function template"
  • Translation: "Translate this passage into English", "Translate into Chinese"
  • Summarization: "Summarize the core viewpoints of this article"
  • Q&A dialogue: "What is Cloud Development", "Explain the mini program lifecycle"

Out of scope

  • Visual requests such as image generation and image editing → not within the scope of this skill, handled by image-gen-skill / image-edit-skill
  • Real-time information queries that require internet search → not within the scope of this skill
  • Operations that require calling specific business APIs (such as ordering, queuing) → not within the scope of this skill

Interface chain

  • generateText: generate text content based on the prompt, supporting system prompts, model selection, and temperature parameters

Usage order

  1. When the user expresses an intent to generate text, directly call generateText
  2. If the user does not specify a model, use the cloudbase default model
  3. If the user specifies a scenario (writing an article/writing code/translation), the AI ​​automatically selects the appropriate systemPrompt and model
  4. After returning the result, display it via the text-result-card card; the user can copy the result, regenerate, or retry with a different model

Design constraints

  • Text generation does not go through cloud functions; directly call the wx.cloud.extend.AI mini program client API
  • Supports three models: cloudbase (general, default), deepseek-v4 (deep reasoning), hunyuan (Chinese-optimized)
  • The system prompt is automatically constructed according to the scenario; when the user does not pass one, it defaults to the general assistant role