bytedance/deer-flow
22 skills45.3k installs2M starsGitHub
Install
npx skills add https://github.com/bytedance/deer-flowSkills in this repo
1Ppt GenerationThe ppt-generation skill Use this skill when the user requests to generate, create, or make presentations (PPT/PPTX). Creates visually rich slides by generating images for each slide and composing them into a PowerPoint file. This skill generates professional PowerPoint presentations by creating AI-generated images for each slide and composing them into a PPTX file. The workflow includes planning the presentation structure with a consistent visual style, generating slide images sequentially (using the previous slide as a reference for style consistency), and assembling them into a final presentation. - Plan and structure multi-slide presentations with unified visual style - Support multiple presentation styles: Business, Academic, Minimal, Apple Keynote, Creative - Generate unique AI images for each slide using image-generation skill - Maintain visual consistency by using previous slide as reference image - Compose images into a professional PPTX file Choose one of the following styles when creating the presentation plan: Style Description Best For ------- ---3.8kinstalls2Claude To DeerflowClaude to DeerFlow is an integration skill for calling a running DeerFlow AI agent platform over HTTP. DeerFlow sits behind an Nginx proxy with a Gateway API for models, skills, memory, and uploads and a LangGraph-compatible API for threads and streaming runs. Environment variables DEERFLOW_URL, DEERFLOW_GATEWAY_URL, and DEERFLOW_LANGGRAPH_URL must be resolved before any curl call, defaulting to localhost port 2026. Primary flow creates a thread, then POSTs a streaming run with assistant_id lead_agent, human messages, stream_mode values and messages-tuple, and context flags for thinking, plan mode, and subagents. Modes range from flash without planning to ultra with subagents enabled. Additional operations list models, skills, agents, memory, upload PDF or Office files converted to markdown, fetch thread history, and search threads. A chat.sh helper checks health, streams SSE events, and prints the final AI message. Errors surface when health fails or the instance is not running.3.3kinstalls3Frontend Designfrontend-design is an agent skill for building distinctive, production-grade web interfaces that avoid generic AI aesthetics. It starts with design thinking around purpose, tone, constraints, and differentiation before committing to a bold aesthetic direction such as brutalist, editorial, retro-futuristic, or refined minimal layouts. Output mandates a working index.html entry file and polished HTML, CSS, or framework code with intentional typography, color systems via CSS variables, motion, spatial composition, and atmospheric backgrounds instead of default purple gradients and overused font stacks like Inter or Roboto. Implementation guidance covers React, Vue, and static layouts with high-impact motion, scroll-triggered reveals, and context-specific visual detail scaled to the chosen aesthetic intensity. A mandatory but subtle Created By Deerflow signature must link to deerflow.tech without competing with main content. Developers reach for it when asked to build landing pages, dashboards, posters, React components, or beautify web UI with creative, memorable design rather than cookie-cutter patterns.3kinstalls4Find SkillsFind Skills helps agents discover and install modular skill packages from the open agent skills ecosystem using the Skills CLI. When a user asks how to do something, requests a skill for a task, or wants to extend agent capabilities, the skill guides keyword search with npx skills find, presents install commands, and links to skills.sh for details. It walks through understanding the domain and task, running targeted searches, presenting options with install paths, and installing chosen skills via install-skill.sh. The skill documents common categories like web development, testing, DevOps, documentation, code quality, design, and productivity with example queries. When no match exists, it acknowledges the gap, offers direct help, and suggests npx skills init for custom skills. Developers reach for it early in a project to avoid reinventing workflows that already ship as installable skills.3kinstalls5Data AnalysisDeer Flow data-analysis analyzes user-uploaded Excel and CSV files with an in-process DuckDB engine through a single analyze.py script. Actions include inspect for schema and sample rows, query for arbitrary SQL including joins and window functions, and summary for per-column statistics on numeric and string fields. Multi-sheet Excel workbooks map each sheet to a table name, CSV files use the filename stem, and multiple files share one query context for cross-file joins. Results export to CSV, JSON, or Markdown tables via --output-file with format auto-detection. The script caches parsed files in a persistent DuckDB database keyed by SHA256 of inputs so repeated inspect-query-summary cycles start near instantly. Documentation covers aggregation patterns, pivot-style CASE sums, CTEs, and date functions on auto-parsed Excel date columns. Agents are instructed to call analyze.py with parameters rather than reading the Python source directly.2.8kinstalls6Video GenerationThe video-generation skill creates AIGC videos using structured JSON prompts and a bundled Python generate.py script. Workflow step one gathers subject, style, technical specs, aspect ratio, and optional reference images without scanning user folders under /mnt/user-data. Step two writes a descriptive JSON prompt file to /mnt/user-data/workspace/. Step three optionally creates a reference image via the image-generation skill; a single image guides the first or last frame. Step four runs generate.py with --prompt-file, optional --reference-images, --output-file, and --aspect-ratio defaulting to 16:9. The skill forbids reading the Python source; call it with parameters only. Provider selection is automatic: GEMINI_API_KEY enables Gemini Veo; only MINIMAX_API_KEY routes to MiniMax async video with first_frame_image; VIDEO_GENERATION_PROVIDER can force gemini or minimax. Outputs land in /mnt/user-data/outputs/ and should be shared via present_files with brief result notes and iteration offers. Prompts must be English regardless of user language.2.4kinstalls7Image GenerationThis skill generates high quality images using structured prompts and a Python script The workflow includes creating JSON formatted prompts and executing image generation with optional reference images Create structured JSON prompts for AIGC image generation Support multiple reference images for style composition guidance Generate images through automated Python script execution Handle various image generation scenarios character design scenes products etc When a user requests image generation identify Subject content What should be in the image Style preferences Art style mood color palette Technical specs Aspect ratio composition lighting Reference images Any images to guide generation You don t need to check the folder under mnt user data The image generation skill documents workflows prerequisites and usage patterns grounded in its repository SKILL md Agents should follow the documented steps respect safety and permission notes and cite only capabilities described in the source It triggers on phrases matching the skill description and integrates with the agent toolchain for the tasks outlined in the documentation2.3kinstalls8Consulting AnalysisThis skill produces professional consulting grade research reports in Markdown format covering domains such as market analysis consumer insights brand strategy financial analysis industry research competitive intelligence investment research and macroeconomic analysis It operates across two distinct phases 1 Phase 1 Analysis Framework Generation Given a research subject produce a rigorous analysis framework including chapter skeleton per chapter data requirements analysis logic and visualization plan 2 Phase 2 Report Generation After data has been collected by other skills synthesize all inputs into a final polished report The output adheres to McKinsey BCG consulting voice standards The report language follows the output_locale setting default zh_CN for Chinese Strict Adherence Rule All data presented in the report and visualized in charts MUST be derived directly from the provided Data Summary or External Search Findings NO Hallucinations Do not invent estimate or simulate data If data is missing state Data not available rather than fabricating numbers Traceable Sources Every major claim and chart must be traceable back to2.3kinstalls9Github Deep ResearchMulti round research combining GitHub API web_search web_fetch to produce comprehensive markdown reports Round 1 GitHub API Round 2 Discovery Round 3 Deep Investigation Round 4 Deep Dive Broad to Narrow Start with GitHub API then general queries refine based on findings Round 1 GitHub API Round 2 topic overview Round 3 topic architecture topic vs alternatives Round 4 topic issues topic roadmap site github com topic Source Prioritization 1 Official docs repos highest weight 2 Technical blogs Medium Dev to 3 News articles verified outlets 4 Community discussions Reddit HN 5 Social media lowest weight for sentiment The github deep research agent skill provides documented workflows prerequisites triggers and safety guidance from its SKILL md source Agents load it when user requests match the description and follow step by step instructions without inventing capabilities It integrates with standard agent tooling for the tasks inputs outputs and failure modes described in the repository documentation2.2kinstalls10Deep ResearchThis skill provides a systematic methodology for conducting thorough web research Load this skill BEFORE starting any content generation task to ensure you gather sufficient information from multiple angles depths and sources Research Questions User asks what is X explain X research X investigate X User wants to understand a concept technology or topic in depth The question requires current comprehensive information from multiple sources A single web search would be insufficient to answer properly Content Generation Pre research Creating presentations PPT slides Creating frontend designs or UI mockups Writing articles reports or documentation Producing videos or multimedia content Any content that requires real world information examples or current data Never generate content based solely on general knowledge The quality of your output directly depends on the quality and quantity of research conducted beforehand A single search query is NEVER enough Start with broad searches to understand the landscape2.1kinstalls11Chart VisualizationThis skill provides a comprehensive workflow for transforming data into visual charts It handles chart selection parameter extraction and image generation 1 Intelligent Chart Selection Analyze the user s data features to determine the most appropriate chart type Use the following guidelines and consult references for detailed specs Time Series Use generate_line_chart trends or generate_area_chart accumulated trends Use generate_dual_axes_chart for two different scales Comparisons Use generate_bar_chart categorical or generate_column_chart Use generate_histogram_chart for frequency distributions Part to Whole Use generate_pie_chart or generate_treemap_chart hierarchical Relationships Flow Use generate_scatter_chart correlation generate_sankey_chart flow or generate_venn_chart overlap Maps Use generate_district_map regions generate_pin_map points or generate_path_map routes Hierarchies Trees Use generate_organization_chart or generate_mind_map Specialized generate_radar_chart Multi dimensional comparison generate_funnel_chart Process stages generate_liquid_chart Percentage Progress generate_word_cloud_chart Text frequency generate_boxplot_chart or generate_violin_chart Statistical d.2kinstalls12Podcast GenerationThe podcast-generation skill Use this skill when the user requests to generate, create, or produce podcasts from text content. Converts written content into a two-host conversational podcast audio format with natural dialogue. It covers convert any text content articles, reports, documentation into podcast scripts. Key workflows include generate natural two-host conversational dialogue male and female hosts. This skill generates high-quality podcast audio from text content. The workflow includes creating a structured JSON script conversational dialogue and executing audio generation through text-to-speech synthesis. Developers invoke podcast-generation when the task matches the triggers and reference files in SKILL.md for grounded, stepwise execution. Reference files and progressive disclosure keep context focused while preserving concrete commands, configuration fields, and validation checks copied from the upstream documentation. Reference files and progressive disclosure keep context focused while preserving concrete commands, configuration fields, and validation checks copied from the upstream documentation.1.9kinstalls13Skill Creatorskill-creator is the Post-hoc Analyzer Agent in bytedance/deer-flow. After a blind comparator picks winner A or B, this skill unblinds results by reading winner_skill_path, loser_skill_path, and both execution transcripts. It explains why the winner performed better and produces concrete suggestions to improve the losing skill's instructions, triggers, and workflow. Developers reach for skill-creator when running structured skill evaluations and need root-cause analysis instead of guessing which SKILL.md change fixed output quality. Inputs include winner label, skill paths, and transcript paths passed directly in the agent prompt.1.8kinstalls14Web Design Guidelinesweb-design-guidelines is a UI review skill from bytedance/deer-flow, authored by Vercel and versioned 1.0.0 in its manifest. The skill fetches fresh Web Interface Guidelines from the canonical source URL on each run, reads user-specified files or patterns, checks every rule in the fetched guideline set, and reports findings in a terse file:line format. Developers reach for it when asked to review UI, check accessibility, audit design, review UX, or validate a site against interface best practices. The workflow is deliberately deterministic: fetch guidelines, scan targets, enumerate violations. It suits React and broader frontend codebases where guideline drift causes inconsistent components, missing focus states, or accessibility gaps. Because guidelines are pulled live, audits reflect current Vercel recommendations rather than stale local copies, making the skill useful in PR review, pre-release QA, and design-system enforcement passes.1.8kinstalls15Academic Paper ReviewThe academic paper review skill Use this skill when the user requests to review, analyze, critique, or summarize academic papers, research articles, preprints, or scientific publications. Supports comprehensive structured reviews covering methodology assessment, contribution evaluation, literature positioning, and constructive feedback generation. Trigger on queries involving paper URLs, uploaded PDFs, arXiv links, or requests like "review this paper", "analyze this research", "summarize this study", or "write a peer review". Documentation covers workflows, commands, and guardrails agents should follow when users invoke this capability. Key documented areas include Parse and comprehend academic papers from uploaded PDFs or fetched URLs; Generate structured reviews following top-venue review templates; Assess methodology rigor (experimental design, statistical validity, reproducibility); Evaluate novelty and significance of contributions. Reference commands include Claim 1: [Specific claim about contribution or finding]; Evidence: [What evidence supports this claim in the paper]. Use when developers or agents need structured guidance for academic paper review tasks with evidence gr.1.8kinstalls16Surprise MeThe surprise-me Deer Flow skill triggers on surprise me, boredom, or requests for something interesting. It reads available skills, picks one to three, and designs a cohesive mashup deliverable rather than separate demos. Principles juxtapose skills unexpectedly, incorporate user context when known, prioritize visual delight over information density, and present with minimal spoilers. Fallbacks when alone include news-based HTML artifacts, interactive mini-games or generative art pages, or personalized creations from memory. Execution reads full SKILL.md bodies, follows their technical steps, merges outputs, and reveals with a short teaser line. Use when users want creative showcases without specifying exact deliverables.1.8kinstalls17Vercel DeployThe vercel-deploy Deer Flow skill deploys applications and websites to Vercel when users request go-live or preview deployments. It returns preview URLs and claimable production links without requiring local Vercel authentication in the documented flow. Agents handle deploy my app, push live, and preview deployment intents with status reporting. Use for quick Vercel shipping workflows from agent sessions. Deploys to Vercel preview and production targets. Returns preview URL and claimable production link. No local authentication required per skill docs. Handles preview and go-live user intents. Deer Flow deployment workflow integration. Deploy applications to Vercel with preview URLs and claimable production links without local authentication. Structured deliverables and steps from the vercel-deploy skill workflow. User mentions vercel-deploy or related skill triggers from the description. How do I apply vercel-deploy for the workflow described in SKILL.md?.1.7kinstalls18Systematic Literature ReviewThe systematic-literature-review skill is designed for produce structured systematic literature reviews and annotated bibliographies across multiple academic papers. Systematic Literature Review Skill Overview This skill produces a structured systematic literature review (SLR) across multiple academic papers on a research topic. Distinct from academic-paper-review: that skill does deep peer review of a single paper. Invoke when the user requests systematic literature review, survey synthesis, or annotated bibliography across papers.1.6kinstalls19Code DocumentationThe code-documentation skill use this skill when the user requests to generate create or improve documentation for code APIs libraries repositories or software projects Supports README generation API reference documentation inline code comments architecture documentation changelog generation and developer guides Trigger on requests like document this code create a README generate API docs write developer guide or when analyzing codebases for documentation purposes Code Documentation Skill Overview This skill generates professional comprehensive documentation for software projects codebases libraries and APIs It follows industry best practices from projects like React Django Stripe and Kubernetes to produce documentation that is accurate well-structured and useful for both new contributors and experienced developers The output ranges from single-file READMEs to multi-document developer guides always matched to the project's complexity and the user's needs Core Capabilities Generate comprehensive README md files with badges installation usage and API reference Create API reference documentation from source code analysis Produce architecture and design documentation with diagrams Wri.1.3kinstalls20Newsletter GenerationThe newsletter-generation skill produces professional researched newsletters combining curated web sources with original commentary. It follows formats inspired by Morning Brew, The Hustle, TLDR, and Benedict Evans, outputting Markdown ready for email platforms or web publishing. Workflow phases plan audience and topic scope, research multiple sources, draft scannable sections with headlines and links, add analysis, and format daily digests, weekly roundups, or deep-dive briefings. Agents adapt tone for technical, executive, or general readers and support recurring series with consistent branding. Triggers include create a newsletter about X, weekly digest, tech roundup, and curated news about Y requests.1.3kinstalls21Bootstrapbootstrap is a conversational onboarding skill in the ByteDance DeerFlow repository that walks developers through 4 adaptive phases across 5–8 rounds to extract language preference, user identity, AI naming, personality traits, and autonomy rules. The skill reads references/conversation-guide.md and templates/SOUL.template.md, then outputs a dense SOUL.md under 300 words written in English. After user confirmation, bootstrap calls DeerFlow's setup_agent tool to persist SOUL.md and finalize the custom agent directory. Developers reach for bootstrap when creating, initializing, or updating a DeerFlow custom agent—triggers include "bootstrap my agent", "create my SOUL.md", or "set up my AI partner".993installs22Music GenerationThis skill generates a song or instrumental track from a JSON spec describing style, mood, and optional lyrics, calling the MiniMax music generation API and returning an MP3. A developer uses it when a user asks to create background music, theme songs, jingles, or instrumental tracks. It requires a MINIMAX_API_KEY and writes the output to a file.136installs