langchain-aiOfficial·22 Agent Skills Included

deepagents

Build production-ready AI agents with planning, memory, and sub-agents

Builds autonomous AI agents that plan multi-step tasks, delegate work to sub-agents, manage files, and remember context across sessions. Removes the need to hand-wire agent loops, context summarization, tool routing, and human approval steps from scratch. Works with any LLM and ships with ready-made skills for research, SQL querying, code review, and content writing.

28,795|4,043

↗

npx skills add langchain-ai/deepagents --all -g -y

Available:

Routes the AI agent to the correct package and scoped guide for each task, and enforces repository-wide rules on testing, security, and safe code changes.

All Skills in This Repository (22)

Pure Emerald Level Indicators

📦 In Repo

langchain-ai

arxiv-search

Searches arXiv for recent/preprint papers and outputs formatted titles and summaries.

langgraph-docs

Fetches LangGraph Python documentation to answer questions and guide implementation via URLs.

skill-creator

Evaluate skill creation, modification, and validation requests for Deepagents CLI workflows.

web-research

Orchestrate multi-source web research with parallel subagents and cited reports.

textual-screenshot

Capture Textual terminal UI states as SVG screenshots using the headless test harness.

skill-creator

Create, initialize, and validate agent skills with SKILL.md frontmatter and bundled resources.

web-research

Delegates web searches to subagents and synthesizes cited research reports.

deepagents-thread-inspector

Inspect and summarize conversations stored in the local Deep Agents Code SQLite session database.

remember

Captures conversation learnings into persistent AGENTS.md memory or reusable skill files.

data-visualization

Create publication-quality matplotlib and seaborn charts saved as PNG files in headless environments.

cudf-analytics

Analyze CSV and tabular datasets with GPU-accelerated NVIDIA cuDF operations.

gpu-document-processing

Extract text, tables, and embeddings from large PDFs using GPU-accelerated sandbox processing.

Official

Intermediate

Frequently Asked Questions

FAQPage Schema

How to install Deep Agents?▼

Run `npx skills add langchain-ai/deepagents --all -g -y` in your terminal to install all skills in this suite globally.

What is Deep Agents used for?▼

It is an open-source agent harness that gives any LLM built-in planning, file access, sub-agent delegation, and long-term memory so it can complete long multi-step tasks on its own.

Does Deep Agents work with models other than OpenAI?▼

Yes. It works with any model that supports tool calling, including Anthropic, Google, open-weight models, and local models via Ollama or vLLM.

How is Deep Agents different from LangChain and LangGraph?▼

LangGraph is the runtime and LangChain's create_agent is a minimal harness; Deep Agents adds opinionated defaults like filesystem access, sub-agents, and context management on top, ready to use out of the box.

Can I create my own custom skills for Deep Agents?▼

Yes. The built-in skill-creator guides your agent through scaffolding, writing, and validating new SKILL.md skills that the agent loads on demand.