Skip to content
Nappai Documentation

Sequential Task Agent

This component in Nappai helps you automate complex tasks by creating a smart AI agent and assigning it a sequence of jobs. Think of it as a sophisticated instruction set for your AI assistant. It allows you to define what your assistant should do, how it should do it, and what resources it can use.

Relationship with CrewAI

This component uses the CrewAI library to create and manage your AI agents and their tasks. CrewAI is a powerful tool that allows for advanced AI capabilities. For more information on CrewAI, visit https://github.com/crewai/crewai.

Inputs

  • Role: Give your AI agent a job title (e.g., “Data Analyst,” “Report Generator”).
  • Goal: Clearly state the overall objective of the agent’s work.
  • Backstory: Provide some background information to help the agent understand the context of its tasks.
  • Tools: Select the tools your agent can use to complete its tasks (options will be provided in the dashboard).
  • Language Model (llm): Choose the AI model that will power your agent (options will be provided in the dashboard).
  • Memory: Enable this if you want the agent to remember previous tasks and information (default is ON).
  • Verbose: Enable this for detailed step-by-step reporting of the agent’s actions (default is ON).
  • Allow Delegation: Allow the agent to assign sub-tasks to other agents (default is OFF).
  • Allow Code Execution: Allow the agent to execute code (default is OFF). Use with caution!
  • Agent kwargs: (Advanced) Additional settings for the agent (only use if you know what you’re doing).
  • Task Description: Describe the task the agent needs to perform in detail.
  • Expected Task Output: Specify the desired outcome of the task.
  • Async Execution: Choose whether the task should run in the background (default is OFF).
  • Previous Task: (Optional) Link this task to a previous task to create a sequence of actions.

Outputs

The component produces a sequence of tasks ready for execution by your AI agent. This output is used by other Nappai components to manage and monitor the workflow.

Usage Example

Let’s say you want to create an agent that analyzes sales data and generates a report. You would:

  1. Set the Role to “Sales Data Analyst”.
  2. Set the Goal to “Generate a weekly sales report.”
  3. Select appropriate Tools (e.g., data access tools, report generation tools).
  4. Choose a Language Model.
  5. Describe the Task Description and Expected Task Output in detail.

The component will then create an agent and task ready for execution.

Templates

This component is used in the “Sequential Tasks Agent” template.

  • Sequential Task: Represents a single task within a sequence. Each task requires a description, expected output, and an assigned agent. [Link to Sequential Task documentation (if available)]
  • Sequential Crew: Manages a group of agents working together on a sequence of tasks. [Link to Sequential Crew documentation (if available)]

Tips and Best Practices

  • Clearly define the roles, goals, and tasks for your agents.
  • Start with simple tasks and gradually increase complexity.
  • Carefully select the appropriate tools and language models for your tasks.
  • Monitor the agent’s progress and make adjustments as needed.

Security Considerations

  • If you enable “Allow Code Execution,” ensure that the code you allow your agent to run is secure and trustworthy. Incorrectly configured code could lead to unintended consequences. Use this feature with extreme caution. Only allow code execution from trusted sources.
  • Regularly review the configurations of your agents and tasks to ensure they align with your security policies.