- Home
- Skills
- Agents & Orchestration
- LangGraph Master Suite: 10 Agent Engineering Skills
Works with the AI Agents you already use
LangGraph Master Suite: 10 Agent Engineering Skills
Professional suite for designing, auditing, and deploying secure, persistent LangGraph agents.
$10
LangGraph Master Suite: 10 Agent Engineering Skills
Example session with this skill installed
Design a LangGraph agent for processing legal contracts. I need a router that sends documents to different specialist nodes, an evaluation loop for quality, and a human approval step.
- Read your context and instructions
- Compiled the langgraph master suite
I have designed a deterministic topology for your Legal Contract Processor. This architecture uses a Triage router to prevent unnecessary LLM calls, an Evaluator-Optimizer loop with a hard retry limit, and a Human-in-the-Loop interrupt before the final commit.
Legal Contract Processor Blueprint
graph TD
START --> Triage[Triage Node]
Triage -->|Analyze| Specialist[Legal Specialist Node]
Triage -->|Invalid| END
Specialist --> Evaluator[Quality Evaluator]
Evaluator -->|Fails| Specialist
Evaluator -->|Passes| HumanGate{{"Human Approval (Interrupt)"}}
HumanGate -->|Approved| Commit[Finalize & Save]
HumanGate -->|Rejected| Specialist
Commit --> END
Architectural Specifications
| Component | Implementation Detail |
|---|---|
| Topology | Router + Evaluator-Optimizer + HITL |
| State Schema | TypedDict with messages: Annotated[list, add_messages] and retry_count: int |
| Safety Gate | Fail-closed Pydantic validator on Specialist output |
| Persistence | AsyncPostgresSaver for cross-session human approval |
| Exit Condition | retry_count > 3 forces transition to END with error state |
Next steps
- Run
langgraph-scaffoldto generate the project directory anduvenvironment. - Use
langgraph-node-builderto implement theTriageandSpecialistasync functions. - Configure the
HumanGateusinginterrupt()and theCommand(resume=...)pattern.
Connects securely to your tools. The creator never sees your data.
About this skill
Stop struggling with unpredictable agents and brittle LangGraph implementations. This suite provides 10 specialized engineering modules to design, build, and deploy production-grade agentic systems. It replaces trial-and-error prompting with rigorous engineering patterns, moving from vague ReAct loops to deterministic topologies and secure state management.
What it does
- Architectural Blueprints: Generates Mermaid diagrams and topology designs for routers, evaluators, and map-reduce flows.
- Pure Async Nodes: Builds deterministic nodes using
Commandrouting andSendparallelism without in-place state mutation. - Safety & Security: Implements fail-closed guardrails, PII scrubbing, and human-in-the-loop (HITL) interrupt patterns.
- Persistence & RAG: Configures
AsyncPostgrescheckpointers, cross-thread memory, and corrective self-RAG pipelines. - Enterprise Integration: Deploys graphs via FastAPI SSE streaming or Telegram bots with proper connection pooling.
How it works
- Scaffold: Initialize a modular project structure using
uvwith separate layers for domain logic and graph orchestration. - Build: Define Pydantic state schemas and implement async nodes that preserve
usage_metadata. - Secure: Add deterministic filters and semantic safety gates to intercept prompt injections.
- Test: Run zero-cost
pytest-asynciounit tests with mocked LLMs and trajectory verification.
Frameworks & tools
Built for LangGraph, LangChain, Pydantic v2, FastAPI, and uv. Supports PostgreSQL/SQLite for persistence and aiogram for messaging.
Why this beats prompting it yourself
Generic prompts often lead to "God Nodes," race conditions, and IDOR vulnerabilities where the LLM controls sensitive IDs. This skill enforces strict architectural boundaries and production-ready patterns that standard chat agents overlook.
Use cases
- Migrating a monolithic Python script into a modular, multi-node LangGraph agent.
- Adding enterprise-grade streaming and human-approval gates to an existing RAG pipeline.
- Auditing an agent codebase for silent state bugs, infinite loops, and memory leaks.
- Building a parallelized map-reduce flow for high-volume document analysis.
Known limitations
Requires Python 3.10+ and langgraph. Does not support synchronous database drivers in async node contexts.
Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Agents & Orchestration.
Requires Python 3.10+ and langgraph. Does not support synchronous database drivers in async node contexts. Requires uv for recommended scaffolding.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before buying a skill.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed 5 days ago
- Software
Python