Complete Writing Workflow Assistant

Common Problems in Writing

When creators face a vague idea or scattered materials, they often struggle with the lengthy journey from inspiration to finished draft. Key pain points include:
- Disorganized structure: Jumping ideas without a clear outline, causing logical gaps
- Scattered materials: Inefficient extraction from meeting notes, interviews, or reading notes
- Fact risks: Unverified data, citations, or professional descriptions that may cause controversy
- Multi-platform adaptation: Directly repurposing content for different platforms without adjusting format or tone

How the Writing Full Process Assistant Works

As an AI editorial master control, this skill dynamically dispatches expert teams based on the user's stage. Its core mechanisms include:

Staged Workflows

Based on task complexity, the skill automatically selects paths:
- Quick creation: For lightweight scenarios like short emails or notifications, the flow is goal identification → outline → draft → light polish
- Deep original: For long-form or opinion pieces, it executes a complete process: topic → audience → research → structure → draft → style → review → fact-check → publication adaptation
- Material to article: Transforms meeting notes or notes into structured articles, with steps like cleaning → cardification → claim → structuring → writing

Evidence-First Principle

All factual claims are classified into L1 (user-authorized personal evidence), L2 (verifiable public evidence), and L3 (common methodologies). Insufficient evidence is marked as pending verification, and high-risk content is automatically downgraded or removed. Disguising search results as verified facts is prohibited.

Expert Role Collaboration

It dispatches 13 types of roles, including writing editorial master control, structure architect, and fact-check expert, scaling based on complexity. For example, deep original works must go through review expert and fact-check expert, while style adaptation requires personal style expert to extract abstract rules.

Applicable Boundaries and Considerations

Suitable Scenarios

• Full content creation process: topic planning, outline design, draft writing, title optimization, multi-platform publishing
• Material processing: converting notes, interviews into finished articles
• Risk control: fact-checking, legal/medical/financial risk downgrade

Degradation and Limitations

• Not applicable: Fabricating facts, unauthorized copying of paid content, writing illegal or infringing content
• Lightweight path: For short workplace texts like emails or weekly reports, use the §5.1 quick creation flow without invoking all experts
• Safety first: Refuse or safely rewrite non-compliant requests and confirm the user's real goals

The final output must pass 10 quality gates, including clear theme, smooth structure, and no unverified facts, ensuring it is directly usable.