Skip to main content

​
Using AI Agents

SkillShop’s AI agents are designed to be intuitive and conversational. This guide will help you understand how to interact with agents effectively and get the most out of their capabilities.

​
Accessing Agents

  1. Main Navigation: Click on “Agents” in the SkillShop sidebar
  2. Agent List: Browse available agents in grid or list view
  3. Agent Details: Click on any agent to view its details and start a conversation

​
Agent Interface

The agent interface provides several key areas:
  • Agent description and capabilities
  • Model configuration and settings
  • Usage statistics and performance metrics

​
Starting a Conversation

​
Basic Interaction

Simply type your question or request in the chat interface. Agents understand natural language and can handle complex, multi-part requests. 
"What are the key concepts in machine learning that I should include in a beginner's curriculum?"

​
Best Practices for Agent Interaction

Provide clear context and specific requirements:
  • Good: “Create a quiz on photosynthesis for 8th grade students with 10 multiple choice questions”
  • Better: “Create a quiz on photosynthesis for 8th grade students with 10 multiple choice questions, focusing on the light-dependent reactions, and include explanations for each answer”
Your specific request becomes the {{ message }} variable that agents use to personalize their responses.
Help agents understand your situation:
  • Grade level or audience
  • Subject area and specific topics
  • Learning objectives or goals
  • Time constraints or requirements
Agents maintain conversation context, so you can:
  • Request modifications to previous responses
  • Ask for clarification or additional details
  • Build on previous work iteratively
Each agent has specialized knowledge:
  • Ask curriculum agents about learning progressions
  • Request assessment agents to create rubrics
  • Have content agents generate multimedia materials

​
Understanding Agent Responses

​
Response Components

Agent responses typically include:
  1. Main Content: The primary answer or deliverable
  2. Tool Usage: Visible actions the agent took (searches, data lookups, etc.)
  3. Reasoning: Explanation of the agent’s approach
  4. Suggestions: Recommendations for next steps or improvements

​
Tool Usage Visibility

When agents use tools, you’ll see:
  • Tool Name: Which capability the agent is using
  • Purpose: Why the agent chose this tool
  • Progress: Real-time updates during tool execution
  • Results: Summary of what the tool accomplished
Tool usage is transparent, so you can understand exactly how agents are gathering information and performing tasks.

​
Working with Different Agent Types

​
System Agents

Best for: General questions, navigation help, platform guidanceExample requests:
  • “How do I create a new learning path?”
  • “What’s the difference between skills and topics?”
  • “Show me recent assessments in mathematics”

​
Workflow Agents

Workflow agents orchestrate complex processes involving multiple steps or agents:
1

Sequential Workflows

Execute agents one after another, passing results between them.Example: “Research → Design → Assess” content creation pipeline
2

Scheduled Workflows

Automate recurring educational tasks at specific times.Example: Weekly curriculum review and updates
3

Loop Workflows

Iterate on content until quality standards are met.Example: Refine assessment questions through multiple review cycles
4

Router Workflows

Intelligently delegate tasks to appropriate specialist agents.Example: Route different types of content requests to specialized creators

​
Advanced Features

​
Session Management

  • Conversation History: Agents remember previous interactions within a session
  • Context Preservation: Build on previous work without repeating information
  • Session Continuity: Return to conversations later and continue where you left off

​
Multi-Agent Collaboration

  • Agent Dispatch: Agents can call other agents for specialized tasks
  • Workflow Integration: Combine multiple agents for complex projects
  • Result Synthesis: Agents can integrate outputs from multiple sources

​
Customization Options

  • Model Selection: Choose different AI models for different tasks
  • Temperature Settings: Adjust creativity vs. consistency
  • Tool Configuration: Enable or disable specific capabilities
  • Memory Settings: Control how much context agents retain

​
Troubleshooting Common Issues

Possible causes:
  • Network connectivity issues
  • Agent model temporarily unavailable
  • Request too complex or ambiguous
  • Template variable processing errors
Solutions:
  • Refresh the page and try again
  • Simplify your request
  • Try a different agent or rephrase your question
  • Check if agent has template variable issues (contact admin)
Possible causes:
  • Insufficient context provided
  • Agent misunderstood the request
  • Tool limitations or data availability
  • Template variables not properly configured
Solutions:
  • Provide more specific context and requirements
  • Ask clarifying questions
  • Try rephrasing your request
  • Ensure your message is clear and specific (it becomes the {{ message }} variable)
Possible causes:
  • External service temporarily unavailable
  • Permission or authentication issues
  • Invalid parameters or data
Solutions:
  • Check system status and try again later
  • Verify your permissions and access rights
  • Contact support if issues persist
Possible causes:
  • Agent template processing failed
  • Missing required template variables
  • Invalid system prompt configuration
Solutions:
  • Try rephrasing your request with more context
  • Use a different agent if available
  • Report the issue to system administrators
  • Check agent configuration if you’re the creator

​
Tips for Success

Start Simple

Begin with straightforward requests to understand how each agent works, then gradually increase complexity.

Iterate and Refine

Use follow-up questions to refine and improve agent responses rather than starting over.

Leverage Expertise

Match your requests to agent specializations for the best results and most relevant responses.

Provide Context

The more context you provide, the better agents can tailor their responses to your specific needs.

​
Next Steps

1

Try Different Agents

Experiment with various system agents to understand their unique capabilities and specializations.
2

Explore Agent Tools

Learn about the tools agents use to understand what they can accomplish for you.
3

Create Custom Agents

Once comfortable with system agents, consider creating custom agents for your specific needs.
4

Build Workflows

Explore workflow agents for complex, multi-step educational processes.