mdadul

Apply Clean Code principles (meaningful names, small functions, error handling, TDD, SRP, smells & heuristics) when authoring, refactoring, or reviewing code—especially TypeScript. Use when the user asks for clean code, refactoring, extract function or class, code smell or maintainability review, readable tests, or Robert C. Martin's Clean Code rules.

Detect, name, and remediate code smells using the classic taxonomy (Bloaters, Object-Orientation Abusers, Change Preventers, Dispensables, Couplers) when reviewing, refactoring, or authoring code. Use when the user asks to find code smells, review code for maintainability, identify long methods/large classes/duplicate code/feature envy, or wants refactoring guidance tied to a named smell.

Create and maintain custom Skills with correct SKILL.md metadata, strong trigger-aware descriptions, concise instructions, progressive disclosure, and validation loops. Use when the user asks to write, improve, review, or organize a Skill.

Create and maintain custom subagent definition files for any AI agent platform, with strong descriptions, tight tool scope, and explicit handoff contracts. Use when the user asks to create, improve, or standardize a subagent.

Teach and apply the Chain of Responsibility (CoR) pattern to pass requests along a dynamic chain of handlers where each can process or forward the request. Use when users ask about CoR, Chain of Command, middleware pipelines, sequential validation, request bubbling, or comparing CoR to Decorator, Command, or Mediator.

Teach and apply the Command pattern to encapsulate requests as objects for decoupled invocation, queuing, scheduling, logging, and undo/redo workflows. Use when users ask about Command, Action, Transaction, invoker-receiver decoupling, reversible operations, or comparing Command with Strategy, Chain of Responsibility, or Memento.

Teach and apply the Iterator pattern to traverse collections without exposing their internal structure. Use when users ask about Iterator, custom traversal algorithms, decoupling traversal from collections, parallel iteration, lazy iteration, or comparing Iterator to for-each loops and language-native iterables.

Teach and apply the Mediator pattern to centralize collaboration rules among tightly coupled components and eliminate direct component-to-component dependencies. Use when users ask about Mediator, Intermediary, Controller, UI dialog coordination, reusable components, or comparing Mediator with Observer, Facade, or Chain of Responsibility.

Evaluate LangChain agent trajectories with AgentEvals trajectory matching or LLM-as-judge, optionally log to LangSmith. Use when scoring tool-call sequences vs a reference, catching prompt or tool regressions, pytest LangSmith runs, or the user mentions agentevals, trajectory match, or evaluators.

Organize and run LangChain agent integration tests against real LLM APIs using pytest markers, env-based secrets, structural assertions, cost controls, and optional HTTP record/replay. Use when tests need live models, CI secrets, flaky LLM output handling, or the user mentions integration tests, real API agents, or pytest -m integration.

Unit-test LangChain agents without API calls using GenericFakeChatModel and in-memory checkpointers. Use when writing pytest tests for create_agent flows, multi-turn state, or scripted tool-call sequences; when avoiding live LLM keys; or when the user mentions fake chat models, agent fixtures, or InMemorySaver for tests.

Implement LangChain handoffs — state-driven behavior changes between agents or configurations using tool-based transitions and Command. Use when building multi-stage conversational flows, customer support pipelines, sequential workflows where capabilities unlock after preconditions, or when an agent needs to transfer control to another agent.

Implement the LangChain agent skills pattern — packaging specialized prompts and domain knowledge as on-demand loadable skills via tool calling. Use when building agents that need many specializations without overwhelming context, distributing development across teams, or implementing progressive disclosure of prompts, schemas, and documentation. Covers basic load_skill pattern, dynamic tool registration, hierarchical skills, and reference-aware skills.

Engineer reliable LangChain agents by controlling model context, tool context, and life-cycle context. Use when building or debugging agents with dynamic prompts, message injection, tool filtering, model selection, structured output, or conversation summarization. Covers state (short-term), store (long-term), and runtime context (static config) patterns using middleware.

Add cross-session persistent memory to LangChain agents using LangGraph stores. Use when building agents that remember user preferences, past interactions, or any data that must persist across conversations. Covers store setup (InMemoryStore, PostgreSQL), namespace/key organization, reading and writing in tools, and vector search.

Design and implement LangChain multi-agent systems using subagents, handoffs, skills, and router patterns. Use when building supervisor/worker agents, state-driven handoffs between agents, on-demand skill loading, or routing pipelines. Covers pattern selection, context engineering across agents, sync/async execution, and tool-based coordination.