UpstageAI

Interact with BotMadang (botmadang.org), a Korean-language community platform for AI agents. Post articles, write comments, upvote/downvote, check notifications, and browse submadangs. Use when user asks to post to BotMadang, check agent notifications, engage with the AI bot community, or manage submadang content.

Build pipelines, agents, RAG flows, and full web services by combining Upstage Solar models, embeddings, and document APIs. Use when building, scaffolding, or deploying anything with Upstage — '솔라로 RAG 만들어줘', 'Upstage 웹앱 만들어줘', 'Upstage로 에이전트 짜줘', 'build a RAG with Solar', 'create an Upstage-based app'. Covers Solar Pro3/Pro2/Mini, embeddings, OCR, document parse, information extraction, classification, schema generation, and Agent API. For single-API one-shot calls (just OCR, just classify), prefer the dedicated upstage-<api> skill instead.

Classify documents into user-defined categories using Upstage Document Classification API. Also supports document splitting for multi-document PDFs. Use when user asks to classify document types, sort documents by category, split a PDF containing multiple documents, or route documents by type.

Parse documents (PDF, images, DOCX, PPTX, XLSX, HWP) into layout-aware markdown/HTML with tables, figures, headings, and bounding boxes using Upstage Document Parse API. Use when user asks to convert documents to markdown/HTML, preserve layout/tables, or analyze document structure — '이 PDF를 마크다운으로 변환해줘', '문서 구조 분석해줘', '표/레이아웃 그대로 추출해줘', 'parse this PDF to markdown'. DO NOT use for plain text-only extraction with word coordinates — use upstage-ocr instead. DO NOT use for schema-driven field extraction (specific values like invoice total) — use upstage-information-extraction instead.

Extract specific named fields from documents using Upstage Information Extraction API with custom JSON schemas (sync/async) or prebuilt models for receipts, invoices, waybills, bills of lading. Use when user wants named values like '청구액', '주문번호', invoice total, supplier name — '영수증에서 금액이랑 날짜 뽑아줘', '인보이스 필드 추출해줘', 'extract invoice number and amount', 'pull structured data from receipts'. DO NOT use for plain text extraction without a schema — use upstage-ocr. DO NOT use for full document layout/markdown conversion — use upstage-document-parse. For schema design help, pair with upstage-schema-generation.

Extract plain text with word-level bounding box coordinates from images and scanned documents using Upstage OCR API. Use when user asks to OCR a document, extract raw text from an image/scan, or get text coordinates — '이미지에서 텍스트만 뽑아줘', 'OCR 돌려줘', 'extract text with coordinates'. DO NOT use for layout-aware extraction (tables, figures, markdown/HTML conversion) — use upstage-document-parse instead. DO NOT use for schema-driven field extraction (invoice number, total amount) — use upstage-information-extraction instead.

Generate a JSON schema for structured document information extraction. Use this skill whenever the user asks to generate or create a schema for a document — including phrases like "스키마 생성해줘", "이 문서에서 뭘 추출할지 스키마 짜줘", "generate a schema for this document", or any request to define extraction fields for a document. Also trigger when the user provides file paths or a folder path containing documents and wants a schema generated.

Delegate longer user-facing text generation to Upstage Solar Pro3 while keeping the primary model for planning and tool calls. Use when drafting explanations, reports, summaries, or other long-form responses. Delegation is controlled by session enablement and a token threshold.

Use when implementing the Detective Agent BE FastAPI service with deterministic rule engine, safe public payloads, event processing, SSE, clean code, and structured observability.

Use when implementing the Detective Agent FE investigation desk with natural-language dialogue, BE/SSE integration, accessibility, clean components, and frontend observability.

Use when turning a raw mystery story into a Detective Agent case package through writer, cross-check writer, and editor approval gates before generating DB seeds and visual assets.

UI/UX design intelligence for web and mobile. Includes 50+ styles, 161 color palettes, 57 font pairings, 161 product types, 99 UX guidelines, and 25 chart types across 10 stacks (React, Next.js, Vue, Svelte, SwiftUI, React Native, Flutter, Tailwind, shadcn/ui, and HTML/CSS). Actions: plan, build, create, design, implement, review, fix, improve, optimize, enhance, refactor, and check UI/UX code. Projects: website, landing page, dashboard, admin panel, e-commerce, SaaS, portfolio, blog, and mobile app. Elements: button, modal, navbar, sidebar, card, table, form, and chart. Styles: glassmorphism, claymorphism, minimalism, brutalism, neumorphism, bento grid, dark mode, responsive, skeuomorphism, and flat design. Topics: color systems, accessibility, animation, layout, typography, font pairing, spacing, interaction states, shadow, and gradient. Integrations: shadcn/ui MCP for component search and examples.

Interact with Langfuse and access its documentation. Use when needing to (1) query or modify Langfuse data programmatically via the CLI — traces, prompts, datasets, scores, sessions, and any other API resource, (2) look up Langfuse documentation, concepts, integration guides, or SDK usage, or (3) understand how any Langfuse feature works. This skill covers CLI-based API access (via npx) and multiple documentation retrieval methods.