How does it handle async LangGraph implementations?

The skill identifies blocking synchronous calls within async nodes and ensures that the correct invocation methods are used for the graph type.

Can it detect issues with graph visualization?

It flags missing entry points, unreachable nodes, and Command returns without destinations that often cause visualization or execution failures.

Does this skill help with state mutation errors?

Yes, it specifically identifies instances where code mutates the state directly instead of returning a partial update, which is a common LangGraph bug.

Does it support Pydantic state models?

Yes, it includes checks for Pydantic models used as graph states to ensure they use Annotated reducers correctly to prevent data loss.

Is it useful for production deployments?

Absolutely. It warns against using InMemorySaver in production and checks for proper thread ID configuration for persistent checkpointers.

LangGraph Code Reviewer

Name: LangGraph Code Reviewer
Author: existential-birds

byexistential-birds

•

Collaboration Tools

Reviews LangGraph workflows for state management bugs, anti-patterns, and structural integrity issues.

About

The LangGraph Code Reviewer skill equips Claude Code with deep domain expertise for auditing LangGraph-based AI applications. It helps developers identify critical issues such as direct state mutations, missing reducers for list fields, and improper checkpointer configurations for interrupts. By analyzing StateGraph structures, conditional edges, and async patterns, this skill ensures that multi-agent systems are robust, thread-safe, and production-ready. It is an essential tool for debugging complex workflows and maintaining high code quality in LangChain-powered projects.

Key Features

Detects state mutation anti-patterns and suggests partial return updates
Verifies checkpointer configurations and thread ID usage for persistent state
Audits graph structural integrity, including entry points and unreachable nodes
14 GitHub stars
Validates StateGraph reducers and message type annotations
Identifies async/sync mismatches and blocking calls in graph nodes

Use Cases

Performing pre-merge code reviews for complex multi-agent AI workflows
Debugging runtime state persistence errors and message history issues
Optimizing graph performance by identifying large state updates and recursion limits

About

Key Features

Detects state mutation anti-patterns and suggests partial return updates
Verifies checkpointer configurations and thread ID usage for persistent state
Audits graph structural integrity, including entry points and unreachable nodes
14 GitHub stars
Validates StateGraph reducers and message type annotations
Identifies async/sync mismatches and blocking calls in graph nodes

Use Cases

Performing pre-merge code reviews for complex multi-agent AI workflows
Debugging runtime state persistence errors and message history issues
Optimizing graph performance by identifying large state updates and recursion limits