Quickstart

Get up and running with UluOps in 5 minutes.

Note

UluOps is in active development — APIs are stabilizing but may still change between minor versions for pre-1.0 packages.

Prerequisites

  1. Node.js 18 or newer — the CLI and SDKs are ESM-only and target modern Node.
  2. Run setup — If you haven't already, run npx @uluops/setup (or npx @uluops/setup --signup to create a free account). This wires UluOps into Claude Code and other detected harnesses. See the Setup guide for details.
  3. Set your environment variable (if not using --shell during setup):
bash
export ULUOPS_API_KEY=ulr_your-api-key-here

Use with Claude Code

The fastest way to use UluOps is directly inside Claude Code and other supported harnesses. UluOps ships as slash commands — no SDK setup needed, just run commands in your terminal.

Run a single agent:

bash
/agents:code-validate

This launches the code-validator agent against your current project, scores the result, and tracks it in the UluOps dashboard.

Run a full workflow:

bash
/pipelines:post-implementation

Executes the post-implementation pipeline — multiple validators run in sequence, each gated by score thresholds. Results are saved as a tracked run with fingerprinted issues.

Other useful commands:

bash
/pipelines:ship            # Final gate before shipping (6-phase validation)
/agents:test-review        # Validate test quality and coverage
/agents:security           # Run security audit
/agents:type-safety        # TypeScript type safety beyond tsc
/issues:check-resolved     # Check if open issues were fixed by recent changes

Claude Code commands connect to UluOps via MCP (Model Context Protocol). Run npx @uluops/setup to configure everything automatically — see the Setup guide.

Track & Integrate with the SDK

The SDKs are the secondary, programmatic path — use them to record results from your own tooling, drive UluOps in CI/CD, or build dashboards. If you're working in Claude Code, you can skip this section.

Install the SDK

Choose your SDK based on what you need:

  • @uluops/core — Run agents, commands, workflows, and pipelines against your code (the execution engine)
  • @uluops/ops-sdk — Track validation runs, manage issues, view analytics
  • @uluops/registry-sdk — Manage definitions, models, and versions
bash
npm install @uluops/ops-sdk

The examples below are TypeScript. To run a .ts file directly, use a TypeScript runtime such as tsx (npm install -D tsx, then npx tsx your-script.ts), or compile with tsc first.

Create a Client

typescript
import { OpsClient } from '@uluops/ops-sdk';

const client = new OpsClient({
  apiKey: process.env.ULUOPS_API_KEY,
});

Save a Run

The ops-sdk records validation results — it doesn't execute agents. To run agents from TypeScript, use @uluops/core; or use the Claude Code commands above or the CLI guide.

typescript
const result = await client.runs.save({
  project: 'my-project',
  workflowType: 'post-implementation',
  agents: [
    { name: 'code-validator', score: 85, decision: 'PASS' },
  ],
  recommendations: [
    {
      agent: 'code-validator',
      title: 'Add error handling',
      priority: 'suggested',
    },
  ],
});

console.log(`Run #${result.run.runNumber}: ${result.correlation.newIssues} new issues`);

Note: recommendations is a top-level array, not nested inside each agent. Each recommendation references its source agent by name.

Next Steps