金色时刻 · 15:48
让好资源,触手可及
AGENT SKILL

lindy-deploy-integration

lindy-deploy-integration 是面向效率工具、数据与分析、开发运维的 Agent Skill,来自 jeremylongshore/claude-code-plugins-plus-skills。查看技能说明、操作步骤与使用约定,并通过原项目了解详细用法。

Configure deployment pipelines for Lindy AI integrations. Use when deploying to production, setting up staging environments, or automating agent deployments. Trigger with phrases like "deploy lindy", "lindy deployment", "lindy production deploy", "release lindy agents".

SKILL.md

下载 MD ↓
namelindy-deploy-integration
descriptionConfigure deployment pipelines for Lindy AI integrations. Use when deploying to production, setting up staging environments, or automating agent deployments. Trigger with phrases like "deploy lindy", "lindy deployment", "lindy production deploy", "release lindy agents".
allowed-toolsRead, Write, Edit, Bash(gh:*), Bash(docker:*)
version1.0.0
licenseMIT
authorJeremy Longshore <[email protected]>

Lindy Deploy Integration

Overview

Configure deployment pipelines for Lindy AI agent integrations.

Prerequisites

  • CI pipeline configured (see lindy-ci-integration)
  • Production Lindy API key
  • Deployment target (Vercel, AWS, GCP, etc.)

Instructions

Step 1: Create Deployment Workflow

# .github/workflows/lindy-deploy.yml
name: Deploy Lindy Integration

on:
  push:
    branches: [main]
  workflow_dispatch:

env:
  LINDY_ENVIRONMENT: production

jobs:
  deploy:
    runs-on: ubuntu-latest
    environment: production

    steps:
      - uses: actions/checkout@v4

      - name: Setup Node.js
        uses: actions/setup-node@v4
        with:
          node-version: '20'
          cache: 'npm'

      - name: Install dependencies
        run: npm ci

      - name: Build
        run: npm run build

      - name: Run pre-deploy checks
        run: npm run predeploy
        env:
          LINDY_API_KEY: ${{ secrets.LINDY_PROD_API_KEY }}

      - name: Deploy to Vercel
        uses: amondnet/vercel-action@v25
        with:
          vercel-token: ${{ secrets.VERCEL_TOKEN }}
          vercel-org-id: ${{ secrets.VERCEL_ORG_ID }}
          vercel-project-id: ${{ secrets.VERCEL_PROJECT_ID }}
          vercel-args: '--prod'

      - name: Sync Lindy agents
        run: npm run sync:agents
        env:
          LINDY_API_KEY: ${{ secrets.LINDY_PROD_API_KEY }}

      - name: Notify Slack
        if: success()
        uses: slackapi/slack-github-action@v1
        with:
          payload: |
            {
              "text": "Lindy integration deployed successfully"
            }

Step 2: Create Agent Sync Script

// scripts/sync-agents.ts
import { Lindy } from '@lindy-ai/sdk';
import fs from 'fs';

interface AgentConfig {
  id?: string;
  name: string;
  instructions: string;
  tools: string[];
}

async function syncAgents() {
  const lindy = new Lindy({ apiKey: process.env.LINDY_API_KEY });

  // Load agent configurations
  const configPath = './agents/config.json';
  const configs: AgentConfig[] = JSON.parse(fs.readFileSync(configPath, 'utf-8'));

  for (const config of configs) {
    if (config.id) {
      // Update existing agent
      await lindy.agents.update(config.id, {
        name: config.name,
        instructions: config.instructions,
        tools: config.tools,
      });
      console.log(`Updated agent: ${config.id}`);
    } else {
      // Create new agent
      const agent = await lindy.agents.create({
        name: config.name,
        instructions: config.instructions,
        tools: config.tools,
      });
      console.log(`Created agent: ${agent.id}`);

      // Update config with new ID
      config.id = agent.id;
    }
  }

  // Save updated config
  fs.writeFileSync(configPath, JSON.stringify(configs, null, 2));
}

syncAgents().catch(console.error);

Step 3: Configure Environments

# .github/workflows/lindy-deploy-staging.yml
name: Deploy to Staging

on:
  push:
    branches: [develop]

jobs:
  deploy:
    environment: staging
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v4
      - run: npm ci
      - run: npm run build
      - run: npm run deploy:staging
        env:
          LINDY_API_KEY: ${{ secrets.LINDY_STAGING_API_KEY }}

Step 4: Add Rollback Capability

// scripts/rollback.ts
import { Lindy } from '@lindy-ai/sdk';

async function rollback(agentId: string, version: string) {
  const lindy = new Lindy({ apiKey: process.env.LINDY_API_KEY });

  // Get version history
  const versions = await lindy.agents.versions.list(agentId);
  const targetVersion = versions.find(v => v.id === version);

  if (!targetVersion) {
    throw new Error(`Version ${version} not found`);
  }

  // Restore to previous version
  await lindy.agents.versions.restore(agentId, version);
  console.log(`Rolled back agent ${agentId} to version ${version}`);
}

rollback(process.argv[2], process.argv[3]).catch(console.error);

Output

  • Automated deployment pipeline
  • Agent sync mechanism
  • Environment-specific deployments
  • Rollback capability

Error Handling

Issue Cause Solution
Deploy failed Build error Check build logs
Agent sync failed Invalid config Validate JSON
Rollback failed Version missing Check version history

Examples

Docker Deployment

# Dockerfile
FROM node:20-alpine
WORKDIR /app
COPY package*.json ./
RUN npm ci --production
COPY dist ./dist
ENV NODE_ENV=production
CMD ["node", "dist/index.js"]
# Deploy job
- name: Build and push Docker image
  run: |
    docker build -t my-lindy-app:${{ github.sha }} .
    docker push my-lindy-app:${{ github.sha }}

Resources

Next Steps

Proceed to lindy-webhooks-events for event handling.

来源:jeremylongshore/claude-code-plugins-plus-skills · MIT 许可 · 许可原文 · MD 为来源页保存的历史版本