Deep Agents Python quickstart

Follow the live docs — do not invent an alternate API from memory:

https://docs.langchain.com/oss/python/deepagents/quickstart

Fetch that page (Docs MCP or HTTP) and implement the research-agent shape it shows (create_deep_agent, research system prompt, invoke with a research question like “What is LangGraph?”).

Local setup constraints

Apply these on top of the quickstart (they keep setup minimal and model-agnostic):

Ask which provider/model to use. Showcase that Deep Agents are model-agnostic. Suggested prompt:

Which model should this agent use? Pass a provider:model string — e.g. openai:gpt-5.5, anthropic:claude-sonnet-5, google_genai:gemini-3.5-flash. Default if you're unsure: anthropic:claude-sonnet-5.
We'll use that provider's built-in web search (no separate search API key).
Create a new directory (e.g. deep-agent/) and do all work there — do not pollute the open project.
Do not use Tavily (or any second search vendor). Replace the quickstart's internet_search / Tavily tool with the chosen provider's built-in web search. Look up the current tool shape on that provider's LangChain chat docs (examples as of writing — re-check if needed):

Provider Built-in search tool

Anthropic {"type": "web_search_20260209", "name": "web_search", "max_uses": 5}

OpenAI {"type": "web_search"}

Google {"google_search": {}}

Prefer Anthropic / OpenAI / Google so provider search is available. Only secret: that provider's API key in .env (gitignored). Skip LangSmith tracing unless they ask.
Install deepagents (+ python-dotenv) and the provider package for their model — not tavily-python.
Run the research example, show output, then stop. Point to deep-agents-core / customization / Managed Deep Agents for next steps.

Provider	Built-in search tool
Anthropic	`{"type": "web_search_20260209", "name": "web_search", "max_uses": 5}`
OpenAI	`{"type": "web_search"}`
Google	`{"google_search": {}}`

Files1

1 files · 1.0 KB

Select a file to preview

Overall Score

76/100

Grade

B

Good

Grades are signals, not a certification. Always review a skill yourself before use.

Safety

78

Quality

74

Clarity

82

Completeness

68

Summary

This skill guides an AI agent to scaffold a minimal local Deep Agent in Python by following the official LangChain quickstart. It replaces the quickstart's third-party Tavily search with provider-native web search (Anthropic, OpenAI, or Google), uses environment variables for API keys, and runs a research example. The skill explicitly constrains setup scope to a new isolated directory.

Static Analysis Findings

1 finding

Patterns detected by deterministic static analysis before AI scoring. Hover over any finding code for detailed information and remediation guidance.

Credential Exposure

SEC-020Direct .env File Access

Direct .env file access

SKILL.md.env

85% confidenceCWE-538: Sensitive Info in Externally-Accessible File

Detected Capabilities

file writedirectory creationenvironment variable accesshttp requestpython package installationshell execution

Trigger Keywords

Phrases that agents use to match this skill to user intent.

deep agent quickstartpython deep agent setuplangchain deepagents localbuild research agentagent web search example

Risk Signals

INFO

Direct .env file access referenced in constraint 3 and setup instructions

SKILL.md | Line: 'Only secret: that provider's API key in `.env` (gitignored)'

Referenced Domains

External domains referenced in skill content, detected by static analysis.

docs.langchain.com

Use Cases

Quickly prototype a Deep Agent locally without complex setup
Learn Deep Agents using a minimal working example
Evaluate Deep Agents across different LLM providers before committing to a full implementation
Set up provider-native web search instead of third-party search APIs for faster iteration

Quality Notes

Skill correctly scopes work to a new isolated directory ('deep-agent/'), reducing risk of polluting the user's project
Clear provider model selection with helpful defaults (anthropic:claude-sonnet-5) makes setup model-agnostic
Well-documented table of provider-specific web search tool shapes prevents agent confusion about API compatibility
Explicit instruction to use .env for secrets and git-ignore it follows best practices for local development
Constraint to avoid Tavily and prefer built-in provider search simplifies dependency management
Instructions to 'fetch that page' and follow live docs ensure agent uses current API shapes rather than stale examples
References external dependency (deepagents package, python-dotenv, provider SDKs) but does not validate installation or handle installation failures
No error handling documented for network failures when fetching docs, invalid provider strings, or missing API keys
No guidance on verifying the agent actually runs or interpreting output
Setup is minimal by design but provides little guidance on next steps beyond pointing to documentation links

Model: claude-haiku-4-5-20251001Analyzed: Oct 12, 2026

Reviews

Add this skill to your library to leave a review.

No reviews yet

Be the first to share your experience.

deepagents-python-quickstart