Skip to content
Back to search

langchain-agent-skills

A collection of agent-optimized LangChain, LangGraph and LangSmith skills for AI coding assistants.

Safety scan

A static text scan, not a guarantee. Select the badge to see what was scanned and every finding.

Checked today

Check it deeper on your machine

It has changed since its scan on 11 Nov 2026.

As a Claude Code plugin, install it and then ask Claude Code:

/plugin marketplace add keithmackay/mackayi/plugin install scanskillsafety@mackayiDo a deep scan of https://github.com/soba-labs/langchain-agent-skills with scanskillsafety

Or from a terminal, clone it and run it:

git clone https://github.com/keithmackay/scanskillsafety-skill ~/.claude/skills/scanskillsafetynode ~/.claude/skills/scanskillsafety/cli.cjs --deep 'https://github.com/soba-labs/langchain-agent-skills'

It checks the exact commit: every file, its packages against OSV (api.osv.dev), how it compares with this page's scan (findsafeskills.com receives the repository's address), and, if you agree, Claude reading every file with your own Claude subscription. It never runs the code.

Badge for your README
findsafeskills safety rating

Markdown

[![findsafeskills safety rating](/api/listings/soba-labs--langchain-agent-skills/badge.svg)](/listings/soba-labs--langchain-agent-skills)

HTML

<a href="/listings/soba-labs--langchain-agent-skills"><img src="/api/listings/soba-labs--langchain-agent-skills/badge.svg" alt="findsafeskills safety rating" /></a>

Install

i

Install instructions from the README

As written by the repo's author. findsafeskills has not verified these steps.

Optional (recommended for contributors): create and sync a local environment.

uv venv --python=3.12
uv sync

1. Pick a skill folder in `skills/` and open its `SKILL.md`. 2. Use it directly in your assistant by pasting relevant sections or pointing the assistant to the file path. 3. For development tasks (creating, validating, packaging skills), follow the commands in `AGENTS.md`.

Example: initialize a new skill using the repo tooling.

uv run skills/skill-creator/scripts/init_skill.py <skill-name> --path skills/

What was scanned

The files the safety scan read, as written by the repository's author and shown as plain text. Lines with a finding are marked.

.claude-plugin/marketplace.json · Manifest · 49 lines

View this file on GitHub (read at the repository's current version: the scan did not record which commit it read)

This file has changed since it was scanned, so this is not exactly the text that was scanned, and no line is marked. The findings are listed with the safety scan above.

  1. {
  2. "name": "soba-labs-langchain-agent-skills",
  3. "owner": {
  4. "name": "Soba Labs",
  5. "email": "simon.budziak@sobalabs.ai",
  6. "url": "https://sobalabs.ai/"
  7. },
  8. "metadata": {
  9. "description": "Agent-optimized LangChain, LangGraph and LangSmith skills for AI coding assistants by Soba Labs",
  10. "version": "2.0.0"
  11. },
  12. "plugins": [
  13. {
  14. "name": "langgraph-skills",
  15. "description": "Collection of LangGraph development skills including project setup, agent patterns, state management, and error handling",
  16. "source": "./",
  17. "strict": false,
  18. "skills": [
  19. "./skills/skill-creator",
  20. "./skills/langgraph-project-setup",
  21. "./skills/langgraph-agent-patterns",
  22. "./skills/langgraph-state-management",
  23. "./skills/langgraph-error-handling"
  24. ]
  25. },
  26. {
  27. "name": "langsmith-skills",
  28. "description": "Collection of LangSmith skills for trace analysis, deployment, and production operations",
  29. "source": "./",
  30. "strict": false,
  31. "skills": [
  32. "./skills/skill-creator",
  33. "./skills/langsmith-trace-analyzer",
  34. "./skills/langsmith-deployment"
  35. ]
  36. },
  37. {
  38. "name": "deepagents-skills",
  39. "description": "Collection of Deep Agents skills for complex long-horizon agents with built-in planning, context management, and task delegation",
  40. "source": "./",
  41. "strict": false,
  42. "skills": [
  43. "./skills/skill-creator",
  44. "./skills/deepagents-setup-configuration",
  45. "./skills/deepagents-planning-todos"
  46. ]
  47. }
  48. ]
  49. }
README.md · README · 196 lines

View this file on GitHub (read at the repository's current version: the scan did not record which commit it read)

This file has changed since it was scanned, so this is not exactly the text that was scanned, and no line is marked. The findings are listed with the safety scan above.

  1. <p align="center">
  2. <a href="https://sobalabs.ai/?utm_source=github&utm_medium=referral&utm_campaign=langchain-agent-skills&utm_content=readme-banner"><img src=".github/soba-labs-banner.svg" alt="Soba Labs" width="100%"></a>
  3. </p>
  4. # Agent Skills for LangChain, LangGraph, LangSmith & Deep Agents
  5. [![GitHub stars](https://img.shields.io/github/stars/soba-labs/langchain-agent-skills?style=flat)](https://github.com/soba-labs/langchain-agent-skills/stargazers) [![License: MIT](https://img.shields.io/badge/license-MIT-ba9768?style=flat)](https://github.com/soba-labs/langchain-agent-skills/blob/main/LICENSE) [![Built by Soba Labs](https://img.shields.io/badge/built%20by-Soba%20Labs-132a3c?style=flat)](https://sobalabs.ai/?utm_source=github&utm_medium=referral&utm_campaign=langchain-agent-skills&utm_content=readme-badge)
  6. A comprehensive collection of agent-optimized skills for AI coding assistants working in the LangChain ecosystem. These skills cover the complete development lifecycle from project setup to production deployment, monitoring, debugging, and Deep Agents setup/configuration.
  7. Each skill is a self-contained package with a `SKILL.md` entry point plus optional scripts, references, and templates.
  8. ## Available Skills
  9. | Skill | Description |
  10. | --- | --- |
  11. | [langgraph-project-setup](./skills/langgraph-project-setup/) | Initialize and configure LangGraph projects (structure, `langgraph.json`, env vars, dependencies). |
  12. | [langgraph-agent-patterns](./skills/langgraph-agent-patterns/) | Multi-agent coordination patterns: supervisor, router, orchestrator-worker, handoffs. |
  13. | [langgraph-state-management](./skills/langgraph-state-management/) | State schemas, reducers, persistence, checkpoint inspection, and migration workflows. |
  14. | [langgraph-error-handling](./skills/langgraph-error-handling/) | Retry strategies, LLM-based recovery loops, and human-in-the-loop escalation patterns. |
  15. | [langgraph-testing-evaluation](./skills/langgraph-testing-evaluation/) | Test and evaluate LangGraph agents with unit/integration patterns, trajectory evaluation, LangSmith dataset evals, and A/B comparisons. |
  16. | [langsmith-trace-analyzer](./skills/langsmith-trace-analyzer/) | Fetch, organize, and analyze LangSmith traces for debugging and performance optimization. |
  17. | [langsmith-deployment](./skills/langsmith-deployment/) | Deploy, monitor, and manage LangGraph applications in production (Cloud, Hybrid, Standalone). |
  18. | [deepagents-setup-configuration](./skills/deepagents-setup-configuration/) | Initialize, configure, and validate Deep Agents projects (Python/JavaScript), including middleware, backends, subagents, persistence, and migration guidance. |
  19. | [deepagents-planning-todos](./skills/deepagents-planning-todos/) | Master the `write_todos` tool for task planning and decomposition in Deep Agents, with patterns for research, coding, analysis tasks, and trace visualization. |
  20. | [skill-creator](./skills/skill-creator/) | Guidance for creating and maintaining skills in this repo. |
  21. ## Coverage
  22. **9 production-ready skills** covering the complete LangChain/LangGraph/Deep Agents development lifecycle:
  23. - 🚀 **Project Setup** - Initialize projects with proper structure and configuration
  24. - 🤝 **Multi-Agent Patterns** - Supervisor, router, orchestrator, and handoff patterns
  25. - 💾 **State Management** - Schemas, reducers, persistence, and checkpointing
  26. - 🛡️ **Error Handling** - Retry policies, LLM recovery, and human-in-the-loop
  27. - ✅ **Testing & Evaluation** - Unit/integration tests, trajectory evaluation, LangSmith evaluation, A/B testing
  28. - 🔍 **Trace Analysis** - Debug with LangSmith traces and pattern detection
  29. - 🌐 **Production Deployment** - Cloud, Hybrid, and Standalone deployment with monitoring
  30. - 🧠 **Deep Agents Setup** - Initialize and configure Deep Agents with validated project scaffolding and compatibility guidance
  31. - 📝 **Deep Agents Planning** - Master write_todos for effective task decomposition with patterns, best practices, and trace visualization
  32. Plus the **skill-creator** meta-skill for extending this collection.
  33. ## Quick Start
  34. Optional (recommended for contributors): create and sync a local environment.
  35. ```bash
  36. uv venv --python=3.12
  37. uv sync
  38. ```
  39. 1. Pick a skill folder in `skills/` and open its `SKILL.md`.
  40. 2. Use it directly in your assistant by pasting relevant sections or pointing the assistant to the file path.
  41. 3. For development tasks (creating, validating, packaging skills), follow the commands in `AGENTS.md`.
  42. Example: initialize a new skill using the repo tooling.
  43. ```bash
  44. uv run skills/skill-creator/scripts/init_skill.py <skill-name> --path skills/
  45. ```
  46. ## Install with the skills CLI
  47. Install every skill into Claude Code, Codex, Cursor, OpenCode and other agents with the open
  48. [skills CLI](https://skills.sh):
  49. ```bash
  50. npx skills add soba-labs/langchain-agent-skills
  51. ```
  52. Or install one skill:
  53. ```bash
  54. npx skills add soba-labs/langchain-agent-skills --skill langgraph-agent-patterns
  55. ```
  56. ## Install as Claude Code Plugin
  57. 1. Add the marketplace:
  58. ```
  59. /plugin marketplace add soba-labs/langchain-agent-skills
  60. ```
  61. 2. Install a plugin bundle:
  62. ```
  63. /plugin install langgraph-skills@soba-labs-langchain-agent-skills
  64. ```
  65. Three bundles are available: `langgraph-skills` (project setup, agent patterns, state management, error handling), `langsmith-skills` (trace analyzer, deployment) and `deepagents-skills` (setup and configuration, planning and todos). Each bundle also ships `skill-creator`.
  66. Or use the interactive menu:
  67. ```
  68. /plugin menu
  69. ```
  70. For local development:
  71. ```
  72. claude --plugin-dir ./path/to/langchain-agent-skills
  73. ```
  74. Once installed, Claude Code will automatically use these skills when relevant.
  75. ## Use in Other Repositories
  76. These skills can be shared by copying a skill folder (for example `skills/langgraph-agent-patterns/`) into another repository or a supported assistant skills directory.
  77. ### OpenAI Codex
  78. Codex reads skills from `.agents/skills` in a repo and `~/.agents/skills` globally
  79. ([Codex customization docs](https://learn.chatgpt.com/docs/customization/overview#skills)).
  80. The `SKILL.md` format is identical to Claude Code's, so no conversion is needed.
  81. ```bash
  82. git clone https://github.com/soba-labs/langchain-agent-skills.git
  83. # For one project
  84. mkdir -p .agents/skills && cp -r langchain-agent-skills/skills/* .agents/skills/
  85. # Or for every project
  86. mkdir -p ~/.agents/skills && cp -r langchain-agent-skills/skills/* ~/.agents/skills/
  87. ```
  88. This repository also ships a Codex plugin manifest at `.codex-plugin/plugin.json`, so it can be
  89. installed as a Codex plugin from a local marketplace rather than copied by hand.
  90. ### OpenCode
  91. OpenCode searches the widest set of locations and will find these skills in any of them. Per project
  92. it reads `.opencode/skills/`, `.claude/skills/` and `.agents/skills/`; globally it reads
  93. `~/.config/opencode/skills/`, `~/.claude/skills/` and `~/.agents/skills/`. Either placement above
  94. therefore works unchanged, or:
  95. ```bash
  96. mkdir -p .opencode/skills && cp -r langchain-agent-skills/skills/* .opencode/skills/
  97. ```
  98. ### Cursor
  99. Option 1: Remote rule (GitHub)
  100. - Cursor Settings → Rules → Add Rule → Remote Rule (GitHub)
  101. - Use: `https://github.com/soba-labs/langchain-agent-skills.git`
  102. Option 2: Local installation
  103. ```bash
  104. # Project-level
  105. git clone https://github.com/soba-labs/langchain-agent-skills.git .cursor/skills/agent-skills
  106. # User-level
  107. git clone https://github.com/soba-labs/langchain-agent-skills.git ~/.cursor/skills/agent-skills
  108. ```
  109. Usage: type `/` in Agent chat to search and select skills by name.
  110. ### Other Assistants
  111. If your assistant does not support skills directly, point it at the skill file:
  112. ```
  113. Read skills/langsmith-deployment/SKILL.md for production deployment guidance
  114. Read skills/langgraph-agent-patterns/SKILL.md for multi-agent patterns
  115. ```
  116. Want these skills set up on your team's codebase, or a LangGraph system taken to production? [Talk to Soba Labs](https://sobalabs.ai/contact/?utm_source=github&utm_medium=referral&utm_campaign=langchain-agent-skills&utm_content=readme-inline).
  117. ## Harness Compatibility
  118. The skills themselves are harness-agnostic: one `SKILL.md` with `name` and `description` frontmatter,
  119. optional `scripts/`, `references/` and `assets/`. Claude Code, Codex and OpenCode all read that same
  120. format and all use progressive disclosure, so nothing in `skills/` is specific to one assistant. Only
  121. discovery and distribution differ, and this repository ships all three:
  122. - **Claude Code** reads `.claude/skills/` (project) and `~/.claude/skills/` (global); distribution via
  123. `.claude-plugin/marketplace.json`.
  124. - **Codex** reads `.agents/skills/` (project) and `~/.agents/skills/` (global); distribution via
  125. `.codex-plugin/plugin.json`.
  126. - **OpenCode** reads `.opencode/skills/`, `.claude/skills/` and `.agents/skills/`, plus the global
  127. equivalents, so it is satisfied by either of the above.
  128. The repo root carries `.claude/skills` and `.agents/skills` as symlinks to `skills/`, which means all
  129. three assistants discover the skills when this repository is itself the working project (contributors
  130. get them for free). Windows contributors may need `git config core.symlinks true`.
  131. ## Repository Structure
  132. - `skills/` - 10 skill packages (9 production skills + skill-creator)
  133. - `.claude-plugin/marketplace.json` - Marketplace manifest for Claude Code
  134. - Optional `*.skill` exports - Generated distribution artifacts
  135. - `PLAN.md` - Roadmap and implementation ordering notes
  136. - `AGENTS.md` - Complete development guidelines and workflow
  137. ## Contributing
  138. Community contributions are welcome. We want to build great, comprehensive, and widely used LangChain agent skills together.
  139. 1. Open an issue for bugs, feature ideas, new skills, or improvements. Include context, expected behavior, and reproducible steps when possible.
  140. 2. Fork the repo and create a focused branch for your change.
  141. 3. Follow `AGENTS.md` while implementing:
  142. - Use `uv run` for Python scripts.
  143. - Keep skill instructions concise and move deep details to `references/`.
  144. - Validate skill structure with `uv run skills/skill-creator/scripts/quick_validate.py skills/<skill-name>/`.
  145. 4. Open a PR and include:
  146. - Clear summary of what changed and why.
  147. - Reference to related issue or `PLAN.md` item (if applicable).
  148. - Validation notes (script checks, `quick_validate.py` output, manual verification).
  149. 5. Request review and respond to feedback quickly so we can merge safely.
  150. Keep changes focused and actionable. `AGENTS.md` is the single source of truth for workflow, tooling, and architecture rules.
  151. ## Read more
  152. - [LangChain agent skills for AI coding assistants](https://sobalabs.ai/blog/langchain-agent-skills/?utm_source=github&utm_medium=referral&utm_campaign=langchain-agent-skills&utm_content=readme-read-more): why we built these skills and how to install them.
  153. - [LangGraph agent skills: decisions, not snippets](https://sobalabs.ai/blog/langchain-agent-skills-real-world-example/?utm_source=github&utm_medium=referral&utm_campaign=langchain-agent-skills&utm_content=readme-read-more): the skills tested on a real six-day research agent.
  154. ## Built by Soba Labs
  155. These skills are free and open source under the MIT license, built and maintained by [Soba Labs](https://sobalabs.ai/?utm_source=github&utm_medium=referral&utm_campaign=langchain-agent-skills&utm_content=readme-footer), an AI-native software development agency that builds and operates production agentic systems. We are LangChain Ambassadors; if you need help taking a LangChain, LangGraph or Deep Agents system to production, [talk to Soba Labs](https://sobalabs.ai/contact/?utm_source=github&utm_medium=referral&utm_campaign=langchain-agent-skills&utm_content=readme-footer) or write to [contact@sobalabs.ai](mailto:contact@sobalabs.ai).

Ratings and comments

Sign in to rate this skill.

Sign in

Free to use. If it’s helped, pass it on.