Browse and discover genes — the atomic capabilities that power evolving AI agents.
对文本内容在数据层进行脱敏（非仅视觉）：默认覆盖主体身份、合同主体（公司名称、地址栏、法定代表人/代理人等）、商业秘密与内部数据、技术凭证与系统权限、文件元数据、特定格式隐蔽内容；支持假名化占位、等长遮蔽或擦除；可选 contractCnV1（仅 pseudonymize）：甲/乙方等主体占位（单侧多法人时为【甲方名1】【甲方名2】类推）、合同编号略称、住址槽位与中段遮罩 18 位身份证等，且与内置 contractParty 规则互斥。可附加自定义正则与说明。不发起网络请求。
Government/law link discovery. chinacourt.cn/article/index/id/… list pages: only #articleList (to paginationControl) → /article/detail/YYYY/MM/id/*.shtml; excludes 要闻/right sidebar on same HTML. CAC /wxzw/zcfg/: article hrefs /YYYY-MM/DD/c_*.htm; followPagination adds POST /cms/JsonList for page 2+ (URL stays index_1.htm). PBOC tiaofasi: list→detail prefers detail HTML URL over attachments inside detail; CSRC/SAFE extractors; NDA /zwgk/zcfb/list/→detail only /zwgk/zcfb/ (excludes /zjjd/ /ytdd/). Generic crawl. Pagination, nested, collectPdf. maxTotalFetches caps GETs+JsonList. No filesystem.
基于知识库与 LLM 的合同条款级审查；输出 riskSummaryStructured 与 revisionSuggestions；扩展 extendedReviewReport（法规对照结论、材料完整性局限、基础形式审查；条款归类含必备与实务高频维；权利义务矩阵摘要；非常规条款标记）。高质量模式、变更优先审、分段与汇总去重、JSON 修复。
Local adaptive optimizer: Fitness, S/A/B/C grades, and declarative recommendations from aggregated de-identified metrics (no network, no FS in-WASM). Host companion applies policy. Full product spec: local-adaptive-self-evolution-scheme.md. 本地自适应：根据宿主聚合的脱敏指标计算 Fitness、S/A/B/C 分级与声明式优化建议 JSON；执行与快照由伴生完成。完整方案见同目录 local-adaptive-self-evolution-scheme.md。
genesis-l0-constraint gene (safety)
genesis-web-search gene (search)
Convert a single .docx or .pptx (OOXML) file from base64 to semantic HTML for downstream kb-article-normalize / ingest. No network; no filesystem. Layout and embedded objects are best-effort; complex charts may be omitted with warnings.
RAG answer: user question plus retrieved chunks (with ids) → markdown answer and citations. Host performs retrieval; keys via apiKeys only.
Parse WeChat-style or generic article HTML into structured fields: title, published_at, author, is_original heuristics, body_text, optional tags. No network; host fetches HTML then passes raw_html.
OpenAI-compatible text embeddings: batch strings in, vectors out. Host writes vectors to pgvector/Chroma. Keys via apiKeys only.
test.hybrid.pub.valid gene (test)
爆款钩子主基因：覆盖 100+ 心理学触发公式、场景模板与高转化开场文案库 编排生成/润色/分析/评分/风控/策略六个子基因；内置多平台适配（小红书/抖音/视频号/TikTok 等），覆盖心理学触发公式；支持 native/llm/hybrid 三种模式，可直接输出短视频标题、封面文案、开场前三秒脚本与优化建议，并回传 Arena 指标持续迭代效果。
按业务场景输出策略路由与公式方向建议。
按吸引力、清晰度、平台适配度进行综合评分。
对钩子进行润色改写，支持 native / llm / hybrid 模式。
执行风险与合规检查，过滤高风险表述。
基于心理学模板与平台特征生成原生钩子草稿。
分析钩子结构与心理触发点，返回诊断与优化建议。
诉讼策略两阶段基因：第一阶段输出结构化策略 JSON；第二阶段在 previousStrategyOutput 约束下输出举证方案、质证提纲与代理意见草稿。知识库字段与合同审查基因区分（程序法、证据规则、可选案例）；脱敏由宿主侧完成。
PPT 工作流：refine 根据 existingPptxBase64 大纲与说明调用 LLM，生成 pptxgen 附录；skipAppendixBuild 时仅返回方案供 Node 侧用模板克隆。文书审查 Agent 可在 HTTP 层上传模板并合并。merge 见 ppt-tool-runner.mjs。
Analyze the writing style of reference WeChat articles and generate new articles matching that style on a given topic. Supports style extraction (tone, structure, rhetoric), outline generation, and full article writing via LLM.
Transform text into a multi-scene video using DeepSeek for scene decomposition and Seedance 2.0 for AI video generation. Each scene becomes a real 4-15s video clip with native audio. Only needs a DeepSeek key + Seedance key.
Create and track text-to-video generation jobs through OpenAI's Videos API. Supports prompt-based generation, optional reference images, and polling until completion.
_test_hybrid_pub_native gene (test)
_test_hybrid_pub_valid gene (test)
Extracts and categorizes all URLs from text content, with optional validation of link accessibility.
Extracts key sentences from text to produce a concise summary within a configurable word limit.
Translate and localize content between languages. Ensure cultural appropriateness and maintain original meaning. Use when translating documents, adapting content for different markets, or creating bilingual content.
Normalizes Markdown formatting: fixes heading levels, standardizes list markers, wraps long lines, and enforces consistent spacing.
Validates JSON data against a JSON Schema definition, returning detailed validation errors with paths.
Lightweight web search returning a single best-match answer with source attribution.
Validate a Gene against L0 safety constraints including memory limits, fuel budgets, and host function permissions.
Format source code according to language-specific conventions. Supports TypeScript, JavaScript, JSON, Markdown, and Rust.
Analyzes source code complexity using cyclomatic complexity, lines of code, and nesting depth metrics.
Maps source file paths to human-readable titles and browsable URLs. Pure mapping logic, no network dependency.
Retrieves relevant documentation chunks via vector similarity search. Computes embeddings and queries Supabase match_documents RPC.
Synthesizes a structured answer from retrieved document chunks using an LLM. Provider-agnostic: supports Claude and OpenAI via environment configuration.
N-body gravitational simulation using spatial hash grid. Partitions particles into cells based on bounding box; each particle interacts only with neighbors in its 3x3 cell neighborhood. Reduces pairwise evaluations for spatially clustered distributions.
N-body gravitational simulation using Barnes-Hut quadtree approximation. Groups distant particles into tree nodes for O(n log n) force computation per step. Balances accuracy and speed via configurable opening angle theta.
N-body gravitational simulation using brute-force all-pairs direct summation. Computes exact O(n^2) pairwise interactions per step — maximum accuracy baseline. Supports preset configurations (solar system, binary star, galaxy cluster, collision).
Conway's Game of Life simulator using bitwise-packed rows. Stores 32 cells per integer for cache-friendly bulk neighbor counting. Optimized for dense, high-throughput simulations on medium-to-large grids.
Conway's Game of Life simulator using sparse set representation. Only tracks live cells and their neighbors, skipping empty regions entirely. Excels on large grids with low population density.
Conway's Game of Life simulator. Computes cellular automaton evolution from an initial grid state over N generations. Supports preset patterns (glider, pulsar, random) and custom grids. Pure computation — ideal Native Gene benchmark.
Intelligent skill orchestration hub. Automatically identifies and dispatches the most suitable skills based on user intent. Analyzes conversation context, matches task types, and combines multiple skills for complex tasks.
Analyze project requirements and recommend suitable open source licenses (MIT, Apache, GPL, AGPL, LGPL, MPL, BSL, SSPL, etc.). Use when the user asks about licensing, choosing a license, open source strategy, or mentions license names like MIT/GPL/Apache.
Creating algorithmic art using p5.js with seeded randomness and interactive parameter exploration. Use this when users request creating art using code, generative art, algorithmic art, flow fields, or particle systems. Create original algorithmic art rather than copying existing artists
Design and optimize AI prompts for maximum effectiveness and cost efficiency. Create structured outputs, few-shot examples, and prompt templates. Use when writing prompts, designing AI features, optimizing token usage, or when the user mentions prompt, system message, or AI output format.
Central project review hub coordinating all skills including product-manager. Perform 360-degree review covering product documentation, architecture, UI, security, performance, testing, and more. Output structured reports with CRITICAL/WARNING/SUGGESTION feedback. Use when reviewing code, PRDs, or when the user asks for project review, code review, architecture assessment, or full audit.
World-class product manager assistant for Web2/Web3/AI domains. Writes PRDs, user stories, feature specs, and product roadmaps. Performs requirements analysis and prioritization.
Guide users through a structured workflow for co-authoring documentation. Use when user wants to write documentation, proposals, technical specs, decision docs, or similar structured content. This workflow helps users efficiently transfer context, refine content through iteration, and verify the doc works for readers. Trigger when user mentions writing docs, creating proposals, drafting specs, or similar documentation tasks.
Write technical documentation including README, API docs, and changelogs. Create clear, structured documentation for AI and Web3 projects. Use when writing documentation, creating README, or when the user mentions docs, documentation, readme, or api docs.
Standardize Git workflow with conventional commits, PR templates, branch naming, and changelog generation. Use when committing code, creating PRs, managing releases, or when the user mentions git, commit, pull request, branch, or changelog.
Automate CI/CD pipelines, Docker configuration, and deployment for Next.js AI/Web3 apps. Setup GitHub Actions, Vercel, and environment management. Use when setting up deployment, CI/CD, or when the user mentions deploy, docker, github actions, vercel, or ci/cd.
Optimize web performance including Core Web Vitals, bundle size, and runtime efficiency. Analyze and improve LCP, FID, CLS metrics. Use when optimizing performance, reducing bundle size, or when the user mentions slow, performance, loading, lighthouse, or web vitals.
Systematic debugging guide for AI and Web3 applications. Diagnose errors, analyze logs, and identify root causes. Use when troubleshooting issues, analyzing errors, or when the user mentions bug, error, not working, debug, or fix.
Audit code for security vulnerabilities in AI and Web3 applications. Check for prompt injection, private key exposure, input validation, and common attack vectors. Use when reviewing security, auditing code, or when the user mentions security, vulnerability, attack, or protection.
Enforce code consistency, naming conventions, and hygiene for AI and Web3 projects. Standardize AI SDK usage, Web3 patterns (wagmi/ethers), and TypeScript types. Use when refactoring code, reviewing for consistency, or when the user mentions code quality, standards, or linting.
Design testing strategies for AI and Web3 applications. Create test cases, mock data, and CI configurations. Use when writing tests, setting up test infrastructure, or when the user mentions testing, test coverage, unit test, integration test, or e2e.
Design database schemas with Prisma for AI and Web3 applications. Define models, relations, indexes, and migrations. Use when designing databases, creating Prisma schemas, or when the user mentions database, schema, model, prisma, or migration.
Design RESTful and tRPC APIs with consistent patterns, error handling, and documentation. Define endpoint structure, response formats, and error codes. Use when designing APIs, creating endpoints, or when the user mentions API, endpoint, REST, tRPC, or response format.
Ensure code has sound business logic, data flow, and technical implementation for AI and Web3 projects. Focus on architecture, edge cases, state management, agent patterns, and wallet safety. Use when writing core logic, reviewing architecture, implementing APIs, or when the user mentions logic, architecture, data flow, or asks for code review ignoring UI.
Execute continuous AI-powered code generation with intelligent iteration loops. Always runs in iterative mode - generates code, validates with quality skills (testing-strategist, tech-lead, logic-architect, debugger), and repeats until success or limits reached. Use when user wants to implement features, write code, or automate programming tasks.
Optimize content for search engines. Improve keyword usage, meta descriptions, headings, and content structure for better ranking. Use when creating web content, blog posts, or any content intended for search discovery.
Verify facts, claims, and sources in written content. Ensure accuracy and credibility of information. Use when writing news, reports, or any content where factual accuracy is critical.
Manage academic citations and references. Format citations according to APA, MLA, Chicago, GB/T 7714 and other styles. Use when writing academic papers or any content requiring proper source attribution.
Check content for plagiarism and ensure originality. Identify potential copied content and guide proper attribution. Use when writing academic papers, articles, or any content requiring original work.
Analyze and adjust content tone to match target audience and purpose. Ensure appropriate voice, formality, and emotional impact. Use when writing for specific audiences or when tone consistency is critical.
Analyze and improve content readability. Evaluate sentence complexity, word choice, and overall accessibility. Use when ensuring content is appropriate for target audience reading level.
Optimize writing style for clarity, consistency, and readability. Improve sentence structure, word choice, and flow. Use when polishing written content for better reader experience.
Check grammar, spelling, and punctuation in written content. Identify common errors and provide corrections. Use when reviewing any written content for grammatical accuracy.
Translate and localize content between languages. Ensure cultural appropriateness and maintain original meaning. Use when translating documents, adapting content for different markets, or creating bilingual content.
Write compelling marketing copy, social media content, and advertising materials. Master platform-specific styles and persuasion techniques. Use when creating content for WeChat, Xiaohongshu, Douyin, or other marketing channels.
Write creative content including stories, scripts, and narrative works. Master storytelling techniques, character development, and emotional expression. Use when creating fiction, screenplays, or any narrative-driven content.
Write professional business documents including reports, proposals, and summaries. Create clear, data-driven content for workplace communication. Use when writing work reports, annual summaries, business plans, or OKRs.
Write academic papers, research reports, and scholarly content. Follow academic conventions, citation styles, and formal writing standards. Use when writing thesis, dissertations, journal articles, or research documents.
Execute continuous AI-powered content generation with intelligent iteration loops. Analyzes writing tasks, selects appropriate skills (grammar-checker, style-optimizer, domain-specific writers), and iterates until quality standards are met. Use when user wants to write documents, articles, reports, or any text content.
Web3 interface component library. Includes wallet connection, network switching, transaction status, and gas estimation components for DApp applications.
AI interface component library. Includes streaming text, chat bubbles, thinking indicators, reasoning display, and token usage components for AI conversation applications.
Unified UI/UX design entry point. Dispatches design tokens, components, and UX pattern sub-skills via style dimension parameters. Supports presets, fine-tuning, and full customization.
Parameterized brand style system. Generates infinite style combinations through 6 dimensions: color temperature, border-radius, motion, density, emotion, and contrast.
Parameterized UI component library. Dynamically generates buttons, inputs, cards, and other component styles based on style parameters.
Parameterized UX interaction pattern library. Generates contextual copy and motion effects based on emotional tone and motion intensity parameters.
Parameterized design token system. Dynamically generates CSS variables for colors, border-radius, spacing, and typography based on brand parameters.
Read file contents from the local filesystem with configurable encoding. Returns content, size, and encoding metadata.
Search the web and return structured results with titles, URLs, and snippets. Supports configurable result count.