Majia Prompt Master Python Edition

Problem It Solves

This Python-based prompt assistant is not a simple translator. It helps turn rough requirements into actionable prompts. It is useful when:
- A user has a general direction but lacks subject, style, scenario, or constraints.
- A task needs structured prompts for image, video, audio, or code generation.
- Intake, safety checks, and mode routing are needed before generation.
- A returning user can reuse preferences in the same domain and avoid repeating context.

How It Works

The workflow centers on intent grading, mode routing, safety checks, and caching:
- Level-1 vague requests trigger M1 research; level-2 requests confirm key details; level-3 requests can be generated directly.
- Routing covers M1 exploratory research, M2 rule reuse, M3 creative expansion, M4 analysis reconstruction, and M5 workflow planning.
- Output passes a confidence gate marked as PASS, WARN, or BLOCK.
- Safety rules are tiered: prohibited content is refused; copyright or portrait issues are substituted; medical or life-related topics are handled with caution.
- Model access is configured in config.yaml for OpenAI, Claude, DeepSeek, Zhipu GLM, Moonshot, or a custom API.
- Local caching stores preferences, allowing a faster path for returning users in the same domain and reducing repeated context length.

Boundaries And Notes

• It is better suited for prompt refinement, generation assistance, and plan decomposition, not commercial responsibility for generated output.
• It applies restrictions to prohibited content, copyright or portrait rights, and medical advice.
• Results depend on the user-supplied API model and confirmed details; complex tasks may still require multiple clarification rounds.