Custom agent logic in LangChain

Introduction

Custom agent logic lets you control how an AI agent thinks and acts step-by-step. It helps you make the agent smarter and fit your special needs.

When you want the agent to follow a specific plan or rules.

When you need the agent to handle special tasks or data formats.

When you want to change how the agent decides what to do next.

When you want to add custom checks or stop conditions during the agent's work.

When you want to combine multiple tools or APIs in a unique way.

Syntax

LangChain

from langchain.agents import Agent

class CustomAgent(Agent):
    def plan(self, intermediate_steps, **kwargs):
        # Your custom planning logic here
        return plan

    def run(self, input):
        # Your custom run logic here
        return output

You create a new class that inherits from Agent.

Override methods like plan or run to change behavior.

Examples

This example makes the agent always choose the 'search' tool with a fixed query.

LangChain

class CustomAgent(Agent):
    def plan(self, intermediate_steps, **kwargs):
        # Always pick the first tool
        return {'tool': 'search', 'input': 'example query'}

This example changes the input before running the normal agent logic.

LangChain

class CustomAgent(Agent):
    def run(self, input):
        # Add a prefix to input before running
        modified_input = 'Check: ' + input
        return super().run(modified_input)

Sample Program

This program creates a custom agent that always uses a 'calculator' tool. The tool just reverses the input string as a fake calculation. When run with '12345', it returns '54321'.

LangChain

class CustomAgent:
    def plan(self, intermediate_steps, **kwargs):
        # Simple plan: always use tool 'calculator' with input
        return {'tool': 'calculator', 'input': kwargs.get('input', '')}

    def run(self, input):
        plan = self.plan([], input=input)
        tool = plan['tool']
        tool_input = plan['input']
        # Simulate tool execution
        if tool == 'calculator':
            # Just return the input reversed as a fake calculation
            return tool_input[::-1]
        return 'No tool executed'

agent = CustomAgent()
result = agent.run('12345')
print(result)

Important Notes

Custom agents let you control the agent's thinking steps.

Override only the methods you need to change.

Test your custom logic carefully to avoid infinite loops or errors.

Summary

Custom agent logic means writing your own rules for how the agent plans and acts.

This helps make agents fit special tasks or workflows.

It involves subclassing and overriding key methods like plan and run.

Practice

1. What is the main purpose of creating custom agent logic in Langchain?

easy

A. To define specific rules for how the agent plans and acts

B. To change the user interface of the agent

C. To improve the speed of the Langchain library

D. To add new data sources automatically

Solution

1. Step 1: Understand the role of custom agent logic

   Custom agent logic is about controlling how the agent decides what to do next.

2. Step 2: Identify the main purpose

   It is used to write your own rules for planning and acting, not for UI or speed improvements.

3. Final Answer:

   To define specific rules for how the agent plans and acts -> Option A

4. Quick Check:

   Custom agent logic = planning and acting rules [OK]

Hint: Custom logic controls agent decisions, not UI or speed [OK]

Common Mistakes:

• Thinking it changes the user interface
• Assuming it speeds up the library
• Believing it adds data sources automatically

2. Which method should you override to customize how an agent decides its next action in Langchain?

easy

A. run

B. plan

C. initialize

D. execute

Solution

1. Step 1: Identify key methods for custom logic

   Langchain agents use methods like plan and run for behavior.

2. Step 2: Determine which controls decision making

   The plan method decides the next action, so overriding it customizes decisions.

3. Final Answer:

   plan -> Option B

4. Quick Check:

   Decision method = plan [OK]

Hint: Override plan() to change agent's next action [OK]

Common Mistakes:

• Overriding run() instead of plan() for decisions
• Using initialize() which is not for planning
• Confusing execute() with plan()

3. Given this custom agent code snippet, what will be printed?

class MyAgent:
    def plan(self, input_text):
        if 'hello' in input_text.lower():
            return 'Greet'
        return 'Ignore'

agent = MyAgent()
print(agent.plan('Hello there!'))

medium

A. Greet

B. Ignore

C. hello

D. Error

Solution

1. Step 1: Analyze the plan method logic

   The method checks if 'hello' is in the input text (case-insensitive). If yes, returns 'Greet'.

2. Step 2: Apply input to the method

   Input is 'Hello there!', which contains 'hello' ignoring case, so it returns 'Greet'.

3. Final Answer:

   Greet -> Option A

4. Quick Check:

   Input contains 'hello' -> returns 'Greet' [OK]

Hint: Check string contains 'hello' ignoring case [OK]

Common Mistakes:

• Ignoring case sensitivity in string check
• Expecting the method to print input text
• Assuming method returns 'Ignore'

4. What is wrong with this custom agent code?

class CustomAgent:
    def plan(self, input_text):
        if input_text.contains('test'):
            return 'Found'
        return 'Not Found'

medium

A. Missing return statement in plan method

B. plan method should not take input_text parameter

C. Indentation error in class definition

D. Using .contains() method which does not exist in Python strings

Solution

1. Step 1: Check string method usage

   Python strings do not have a .contains() method; use 'in' keyword instead.

2. Step 2: Identify correct string membership check

   Correct way is: 'test' in input_text, so .contains() causes an error.

3. Final Answer:

   Using .contains() method which does not exist in Python strings -> Option D

4. Quick Check:

   Python strings use 'in', not .contains() [OK]

Hint: Use 'in' keyword for substring check in Python [OK]

Common Mistakes:

• Using .contains() instead of 'in'
• Thinking plan method can't take parameters
• Assuming missing return statement

5. You want to create a custom Langchain agent that first plans actions based on input, then logs each action before running it. Which approach correctly combines planning and running with logging?

hard

A. Override only run() to plan, log, and execute actions

B. Override plan() to log actions and run() to decide actions

C. Override plan() to decide actions and override run() to log and execute actions

D. Use default plan() and run(), add logging outside the agent

Solution

1. Step 1: Understand responsibilities of plan() and run()

   plan() decides what to do next; run() executes actions.

2. Step 2: Combine logging with correct methods

   Override plan() for custom decisions; override run() to add logging before execution.

3. Final Answer:

   Override plan() to decide actions and override run() to log and execute actions -> Option C

4. Quick Check:

   plan() decides, run() logs and executes [OK]

Hint: Plan decides; run executes and logs actions [OK]

Common Mistakes:

• Putting planning logic inside run() only
• Logging inside plan() without running actions
• Not overriding methods and logging externally