jenkins

🌟 Features

• 🧩 Single client, full surface — JenkinsClient covers job/build metadata, console logs (with live streaming), artifacts, the build queue, and triggers
• ⚙️ Layered config resolution — named server → environment variables → per-call overrides, so callers never hardcode credentials
• 🗂 Multi-server profiles — add, switch, and remove named Jenkins servers without touching env vars
• 🔁 Resilient HTTP layer — timeout and bounded retry with backoff on network errors and 5xx responses, automatic CSRF crumb handling on every mutating call
• 🌲 Recursive job search — BFS traversal across nested folders, plus an incremental/concurrent variant with progress callbacks for large instances
• 🧼 Log sanitisation — strips ANSI escape codes and control characters from console output for clean rendering
• 📐 Fully typed — exported TypeScript types for jobs, builds, artifacts, parameters, and changesets

Surface-agnostic by design: this package holds no CLI, no prompts, no rendering — just data, actions, and auth. @kud/jenkins-cli is the command surface and interactive TUI built on top of it.

🚀 Quick Start

npm install @kud/jenkins

import { JenkinsClient, resolveConfig } from "@kud/jenkins"

const cfg = resolveConfig()
const client = new JenkinsClient(cfg.url, cfg.user, cfg.token)

const jobs = await client.searchJobs("my-service")
const build = await client.getBuild("my-service")

console.log(build.result, build.building)

Configuration resolves in layers — named server profile → environment variables → explicit overrides:

export JENKINS_URL=https://ci.example.com
export JENKINS_USER=erwann
export JENKINS_API_TOKEN=xxxxxxxxxxxx

Several instances can be declared at once, comma or pipe separated and matched positionally. JENKINS_INSTANCES names them; without it, each name is derived from its URL's first hostname label. The first instance is the default, and JENKINS_SERVER (or a server override) selects another:

export JENKINS_INSTANCES=pipeline,scheduler
export JENKINS_URL=https://pipeline.example.com,https://scheduler.example.com
export JENKINS_USER=erwann          # a single value applies to every instance
export JENKINS_API_TOKEN=tok1,tok2

A name/URL count mismatch is an error rather than a best-effort pairing: a short list would silently authenticate one instance with another's credentials.

Or manage named servers on disk (~/.config/jenkins-cli/config.json, override the base with XDG_CONFIG_HOME):

import { addServer, useServer, listServers } from "@kud/jenkins"

addServer("prod", {
  url: "https://ci.example.com",
  user: "erwann",
  token: "xxxx",
})
useServer("prod")
listServers() // [{ name: "prod", url: "...", current: true }]