wishket-portfoliolisted

# wishket-portfolio — Wishket Portfolio Draft Generator ## Flow ```mermaid flowchart LR A[Collect inputs: URL/repo/docs/folder] --> B[Analyze materials] B --> C{Missing required info?} C -- Yes --> D[AskUserQuestion] D --> E C -- No --> E[Draft plain-text form] E --> F[Validate: prohibited info / char limits] F --> G[Save to portfolios/ + Summary] ``` ## Step 1: Input Collection Extract context by input type: - **Live URL**: Fetch main pages via WebFetch or browser tools. Identify service purpose, key features, target audience. - **GitHub Repository**: Fetch README and raw manifest files (`package.json`, `Cargo.toml`, `pubspec.yaml`, `requirements.txt`, `go.mod`) to confirm tech stack. - **Requirement Documents (PDF/PPT/DOCX/XLSX)**: - PDF: Read directly. - PPTX: `unzip -p <file>.pptx 'ppt/slides/*.xml' | sed -e 's/<[^>]*>/ /g'`. - DOCX: `unzip -p <file>.docx word/document.xml | sed -e 's/<[^>]*>/ /g'`. - XLSX: `unzip -p <file>.xlsx 'xl/sharedStrings.xml' | sed -e 's/<[^>]*>/ /g'`. - Fallback: Request text copy-paste from user if extraction fails. - **Local Project Folder**: Inspect README, manifests, and directory tree. Identify core architecture and features. Cross-validate multi-source inputs: Code/manifests take precedence for tech stack; requirement docs/websites take precedence for business context. ## Step 2: Verify Required Information Never invent fields that cannot be inferred from code. Inquire via `AskUserQuestion`: -