Project Documentation Generator Plus

The problem

Project planning often remains scattered across long conversations: tech-stack choices, module boundaries, API design, risks, and deployment assumptions are discussed in different turns. Once the planning session ends, it is hard to turn that context into a reviewable engineering document. project-doc-generator-plus addresses this by extracting structured project information from conversation context and producing a first-draft document package.

How it works

The skill first clarifies the analysis scope: current session, recent N messages, a specific topic, or a time range. It then uses sessions_history to collect relevant records and extracts project basics, architecture, functional requirements, database design, API design, integrations, and an inferred directory layout.

The output is engineering-oriented:
- Multi-format documents: Markdown, Word, PPT, Excel, and HTML; Word output can be produced with python-docx, pandoc, or HTML conversion.
- Diagrams and data tables: Mermaid code for architecture, flow, sequence, ER, and Gantt diagrams; Excel sheets for database fields, API lists, module priorities, and risks.
- Analytical extras: tech-stack fit, architecture risk, cost, timeline, and completeness checks for requirements and documentation.
- Review workflows: useful for consolidating team discussions, comparing planning versions, and adding comments before formal submission.

Boundaries and notes

Output quality depends on the conversation. If the discussion lacks interface details, permission models, deployment constraints, or data schemas, the result is better treated as an initial draft. For large projects, generate module-by-module and keep human review for architecture and risk decisions. Optional formats such as Word, PPT, and Excel may require corresponding libraries or tools in the environment.