论文内容智能生成工作流 (Automated Academic Content Generation Pipeline)

What problem does it solve?

This Skill converts provided project code and datasets into a structured academic paper draft, enforcing rigorous academic tone while reducing hallucinations through source-of-truth constraints.

Core Features & Use Cases

• Strict academic prompting: Produces formal, objective CS/SE thesis-style writing with mandated formatting (tables and Mermaid diagrams) and explicit instruction to avoid first-person or casual phrasing.
• Lazy loading & JIT context: Uses directory tree indexing and section-scoped fetching to prevent context overflow in large repositories.
• Outline-to-draft workflow with mandatory user review: Generates an outline first, then blocks execution for user approval before drafting each stage.
• Quality assurance & integrity gates: Enforces physical QA checks (character counts, media presence, TODO/empty section detection), an integrity report, and a multi-role self-review with a score threshold before handoff to the next pipeline.

Quick Start

Ask the agent to generate an academic paper draft from your repository and dataset using the content_generation workflow, then stop at the required outline approval step before writing the full sections.