lindy-core-workflow-a
Core Lindy workflow for creating and configuring AI agents.Use when building new agents, defining agent behaviors,or setting up agent capabilities.Trigger with phrases like "create lindy agent", "build lindy agent","lindy agent workflow", "configure lindy agent".
When & Why to Use This Skill
This Claude skill provides a comprehensive, standardized workflow for creating and configuring Lindy AI agents. It streamlines the entire development lifecycle—from defining agent personas and instructions to integrating tools and performing automated testing—ensuring that AI agents are deployed with consistent behaviors and optimized performance.
Use Cases
- Rapid Prototyping: Quickly build and deploy specialized AI agents (e.g., Customer Support or Sales) by defining specific instructions and toolsets through a structured SDK-based workflow.
- Agent Behavior Optimization: Use the standardized specification format to fine-tune agent personas and instructions, ensuring the AI adheres to professional standards and brand guidelines.
- Automated Tool Integration: Seamlessly connect agents to external tools like email, calendars, or knowledge bases, and verify their functionality through built-in testing scripts.
- Developer Workflow Standardization: Establish a consistent process for teams to create, document, and test AI agents, reducing errors in configuration and deployment.
| name | lindy-core-workflow-a |
|---|---|
| description | | |
| allowed-tools | Read, Write, Edit |
| version | 1.0.0 |
| license | MIT |
| author | Jeremy Longshore <jeremy@intentsolutions.io> |
Lindy Core Workflow A: Agent Creation
Overview
Complete workflow for creating, configuring, and deploying Lindy AI agents.
Prerequisites
- Completed
lindy-install-authsetup - Understanding of agent use case
- Clear instructions/persona defined
Instructions
Step 1: Define Agent Specification
interface AgentSpec {
name: string;
description: string;
instructions: string;
tools: string[];
model?: string;
temperature?: number;
}
const agentSpec: AgentSpec = {
name: 'Customer Support Agent',
description: 'Handles customer inquiries and support tickets',
instructions: `
You are a helpful customer support agent.
- Be polite and professional
- Ask clarifying questions when needed
- Escalate complex issues to human support
- Always confirm resolution with the customer
`,
tools: ['email', 'calendar', 'knowledge-base'],
model: 'gpt-4',
temperature: 0.7,
};
Step 2: Create the Agent
import { Lindy } from '@lindy-ai/sdk';
const lindy = new Lindy({ apiKey: process.env.LINDY_API_KEY });
async function createAgent(spec: AgentSpec) {
const agent = await lindy.agents.create({
name: spec.name,
description: spec.description,
instructions: spec.instructions,
tools: spec.tools,
config: {
model: spec.model || 'gpt-4',
temperature: spec.temperature || 0.7,
},
});
console.log(`Created agent: ${agent.id}`);
return agent;
}
Step 3: Configure Agent Tools
async function configureTools(agentId: string, tools: string[]) {
for (const tool of tools) {
await lindy.agents.addTool(agentId, {
name: tool,
enabled: true,
});
}
console.log(`Configured ${tools.length} tools`);
}
Step 4: Test the Agent
async function testAgent(agentId: string) {
const testCases = [
'Hello, I need help with my order',
'Can you check my subscription status?',
'I want to cancel my account',
];
for (const input of testCases) {
const result = await lindy.agents.run(agentId, { input });
console.log(`Input: ${input}`);
console.log(`Output: ${result.output}\n`);
}
}
Output
- Fully configured AI agent
- Connected tools and integrations
- Tested agent responses
- Ready for deployment
Error Handling
| Error | Cause | Solution |
|---|---|---|
| Tool not found | Invalid tool name | Check available tools list |
| Instructions too long | Exceeds limit | Summarize or split instructions |
| Model unavailable | Unsupported model | Use default gpt-4 |
Examples
Complete Agent Creation Flow
async function main() {
// Create agent
const agent = await createAgent(agentSpec);
// Configure tools
await configureTools(agent.id, agentSpec.tools);
// Test agent
await testAgent(agent.id);
console.log(`Agent ${agent.id} is ready!`);
}
main().catch(console.error);
Resources
Next Steps
Proceed to lindy-core-workflow-b for task automation workflows.