Use when building, wiring, or debugging an Agentex agent — choosing agent type, configuring acp.py and manifest.yaml, using adk.messages or adk.state, or resolving Windows-specific setup issues.
EnglishMachine learning toolkit for big data teams. Includes scikit-learn, PyTorch Lightning, Transformers, SHAP for model training, deployment, and interpretation. Use when building ML pipelines, training models, or explaining predictions.
EnglishCore big data processing toolkit for data teams. Includes Polars, Dask, Vaex for large-scale data processing, ETL pipelines, and distributed computing. Use when working with datasets larger than memory, building data pipelines, or optimizing data processing performance.
EnglishData visualization toolkit for big data teams. Includes Matplotlib, Seaborn, Plotly for static and interactive charts. Use when creating dashboards, reports, or exploring data visually.
EnglishThis skill upgrades reflection from \"append another note\" to a compact learning…
EnglishUse when: decision-making, strategy, risk assessment, problem analysis, growth planning, or when the user asks 'which framework/model should I use'. Trigger: 思维模型, mental model, framework, 该用什么框架, 怎么分析, 决策, 战略, 决策框架, munger, lattice, 格栅. NOT for: direct coding tasks, UI implementation, routine operational work.
EnglishUse when building a RAG pipeline that ingests PDFs, Excel, CSV, or images — especially when debugging silent data loss, choosing between OCR tools, or handling edge cases like scanned pages, merged cells, or embedded charts.
EnglishDocker and Kubernetes workflows for containerization. Use for building containers, orchestration, and cloud-native deployments.
EnglishUse when publishing a 深度学习系列 episode (deep-learning topic deep-dive — interview / paper / framework / debate) end-to-end with editorial polish. Pipeline: signature visual style (one per episode, never reuse) → HTML body → dedicated CF Pages subdomain → triple-format screenshot (4K + mobile + WeChat JPEG) → GDrive structured archive. Trigger: 深度学习管线 / learning pipeline / 专题报告 / 系列报告 / 出一期 / 新一期 / 下一期 / 第N期 / 罗福莉式 / hermes-agent式. NOT for: passive content summary (use deep-learning-skill for body generation; this skill orchestrates the full delivery pipeline around it). Mode: pipeline (sequential A→B→C→D→E).
EnglishUse when choosing a deployment strategy for a release, setting up canary or blue/green rollouts, adding feature flags to decouple deployment from release, coordinating a zero-downtime database migration, or defining rollback criteria and procedures.
EnglishPerform Five Forces analysis — competitive rivalry, supplier power, buyer power, threat of substitutes, and threat of new entrants. Use when analyzing industry dynamics, assessing competitive forces, or evaluating market attractiveness.
EnglishPaper workflow: read papers + cast cards in one go. Takes one or more arxiv links, paper URLs, PDFs, or paper names. For each paper, runs ljg-paper (generates org analysis) then ljg-card -l (generates long reading card PNG). Use when user says '论文流', 'paper flow', '读论文并做卡片', '论文卡片', or provides multiple papers wanting both analysis and cards.
EnglishUse when installing or configuring memorymap, writing CLAUDE.md session-setup instructions, choosing what to save in memory vs history, managing cross-project memory, tuning compression, or troubleshooting why Claude isn't loading context at session start.
EnglishUse when the user wants an 8-expert or multi-expert roundtable, multi-perspective strategic analysis, consulting-style decision report, expert challenge session, pyramid-principle report, or front-door strategic thinking workflow. Always collect both a clear core question and relevant context before analysis.
EnglishUse when searching across both Chinese and international sources simultaneously (17 engines: 8 CN + 9 Global). Trigger: 'search Baidu and Google', 'Chinese + English sources', cross-region market research, privacy-first search, WolframAlpha knowledge queries. NOT for: single-engine queries (use ask-search), real-time finance data, or when speed is critical (parallel requests are slower).
EnglishUse when building or debugging OpenAI Agents SDK workflows — defining agents with tools and handoffs, wiring typed context, streaming responses, adding guardrails, or integrating with the Agentex ADK.
EnglishBenchmark indicator performance with BenchmarkDotNet. Use for Series/Buffer/Stream benchmarks, regression detection, and optimization patterns. Target 1.5x Series for StreamHub, 1.2x for BufferList.
EnglishDesign a product metrics dashboard with North Star metric, input metrics, health metrics, and alert…
English
AGENT SKILLS ATLASIncludes Bigdata Machine Learning, Bigdata Processing, Bigdata Visualization, and other workflows.
Claude Code
Codex
Cursor
GitHub Copilot
Gemini CLISupported by the Atlas install prompt. Skill-specific requirements vary.
Pack stats
What’s included
Agentex
Use when building, wiring, or debugging an Agentex agent — choosing agent type, configuring acp.py and manifest.yaml, using adk.messages or adk.state, or resolving Windows-specific setup issues.
Bigdata Machine Learning
Machine learning toolkit for big data teams. Includes scikit-learn, PyTorch Lightning, Transformers, SHAP for model training, deployment, and interpretation. Use when building ML pipelines, training models, or explaining predictions.
Bigdata Processing
Core big data processing toolkit for data teams. Includes Polars, Dask, Vaex for large-scale data processing, ETL pipelines, and distributed computing. Use when working with datasets larger than memory, building data pipelines, or optimizing data processing performance.
Bigdata Visualization
Data visualization toolkit for big data teams. Includes Matplotlib, Seaborn, Plotly for static and interactive charts. Use when creating dashboards, reports, or exploring data visually.
Cognitive Reflection
This skill upgrades reflection from \"append another note\" to a compact learning system:
Cognitive Skeleton
Use when: decision-making, strategy, risk assessment, problem analysis, growth planning, or when the user asks 'which framework/model should I use'. Trigger: 思维模型, mental model, framework, 该用什么框架, 怎么分析, 决策, 战略, 决策框架, munger, lattice, 格栅. NOT for: direct coding tasks, UI implementation, routine operational work.
Complex Doc RAG
Use when building a RAG pipeline that ingests PDFs, Excel, CSV, or images — especially when debugging silent data loss, choosing between OCR tools, or handling edge cases like scanned pages, merged cells, or embedded charts.
Containerization
Docker and Kubernetes workflows for containerization. Use for building containers, orchestration, and cloud-native deployments.
Deep Learning Pipeline
Use when publishing a 深度学习系列 episode (deep-learning topic deep-dive — interview / paper / framework / debate) end-to-end with editorial polish. Pipeline: signature visual style (one per episode, never reuse) → HTML body → dedicated CF Pages subdomain → triple-format screenshot (4K + mobile + WeChat JPEG) → GDrive structured archive. Trigger: 深度学习管线 / learning pipeline / 专题报告 / 系列报告 / 出一期 / 新一期 / 下一期 / 第N期 / 罗福莉式 / hermes-agent式. NOT for: passive content summary (use deep-learning-skill for body generation; this skill orchestrates the full delivery pipeline around it). Mode: pipeline (sequential A→B→C→D→E).
Deployment Strategies
Use when choosing a deployment strategy for a release, setting up canary or blue/green rollouts, adding feature flags to decouple deployment from release, coordinating a zero-downtime database migration, or defining rollback criteria and procedures.
Ft Business Analyst
Use when: understanding business scenarios, identifying tax-related
Ft Compliance Auditor
Use when: risk identification, compliance checking, Golden Tax IV alerts,
Ft Compliance Checklist
Use when: generating compliance checklists, monthly/quarterly/annual
Ft Risk Assessment
Use when: evaluating tax compliance risk, benchmarking financial indicators
Ft Tax Advisor
Use when: tax type determination, tax rate application, tax filing,
Ft Tax Planner
Use when: client needs tax optimization strategy, comparing tax burden
Industry Forces
Perform Five Forces analysis — competitive rivalry, supplier power, buyer power, threat of substitutes, and threat of new entrants. Use when analyzing industry dynamics, assessing competitive forces, or evaluating market attractiveness.
Ljg Paper Flow
Paper workflow: read papers + cast cards in one go. Takes one or more arxiv links, paper URLs, PDFs, or paper names. For each paper, runs ljg-paper (generates org analysis) then ljg-card -l (generates long reading card PNG). Use when user says '论文流', 'paper flow', '读论文并做卡片', '论文卡片', or provides multiple papers wanting both analysis and cards.
Memory Map
Use when installing or configuring memorymap, writing CLAUDE.md session-setup instructions, choosing what to save in memory vs history, managing cross-project memory, tuning compression, or troubleshooting why Claude isn't loading context at session start.
Multi Expert Roundtable Report
Use when the user wants an 8-expert or multi-expert roundtable, multi-perspective strategic analysis, consulting-style decision report, expert challenge session, pyramid-principle report, or front-door strategic thinking workflow. Always collect both a clear core question and relevant context before analysis.
Multi Search Engine
Use when searching across both Chinese and international sources simultaneously (17 engines: 8 CN + 9 Global). Trigger: 'search Baidu and Google', 'Chinese + English sources', cross-region market research, privacy-first search, WolframAlpha knowledge queries. NOT for: single-engine queries (use ask-search), real-time finance data, or when speed is critical (parallel requests are slower).
OpenAI Agents
Use when building or debugging OpenAI Agents SDK workflows — defining agents with tools and handoffs, wiring typed context, streaming responses, adding guardrails, or integrating with the Agentex ADK.
Performance Testing
Benchmark indicator performance with BenchmarkDotNet. Use for Series/Buffer/Stream benchmarks, regression detection, and optimization patterns. Target 1.5x Series for StreamHub, 1.2x for BufferList.
Pm Cmd Setup Metrics
Design a product metrics dashboard with North Star metric, input metrics, health metrics, and alert thresholds
Showing 24 of 38 skills. Use the list’s page controls above to explore the rest.
Quick start
- Choose your skills. Select individual skills above, or choose the entire pack.
- Copy the install prompt. Use Copy install prompt above the list.
- Send it to your agent. Paste into your coding agent’s chat. Your agent reviews the source and guides the setup.
Try it with your agent
After installation, start with a request like:
Use the Agentex skill to help with my task.
Ask me for the context you need before getting started.Requirements & license
- A coding agent that can access your project files and install skills.
- Your agent checks each skill’s instructions, dependencies, and permissions during setup.
- Reported license for the skills shown: MIT. The author’s terms apply.