go-agent-development_skill

Overview

This skill is a concise Go Agent development guide for building CI/CD build agents. It covers architecture, heartbeat and startup flows, task execution patterns, logging, upgrade handling, and interaction with the Dispatch/Backend. Use it to implement robust, maintainable build agents and supporting libraries in Go.

How this skill works

The guide describes a two-process layout: a DevopsDaemon (watchdog) and DevopsAgent (main process), and outlines core packages such as api, config, job, pipeline and util. It explains common operations: agent startup and heartbeat via HTTP clients, task dispatch and execution via the job package, logging patterns, concurrency with goroutines and channels, and retry/error handling strategies. Examples show HTTP client usage, configuration access, and safe goroutine lifecycle and panic recovery.

When to use it

• When implementing a build machine agent responsible for executing CI jobs in Go.
• When integrating agent lifecycle actions (startup, heartbeat, upgrade) with backend APIs.
• When designing task execution logic, parallelism, and resource cleanup for jobs.
• When adding monitoring, logs, or telemetry collection from an agent host.
• When writing Go utility libraries that interact with the agent runtime or config.

Best practices

• Centralize configuration in a single AgentConfig object and expose safe getters.
• Always log with context (Debug/Info/Error) and include errors via WithError.
• Wrap goroutine entry points with defer recover to avoid process crashes.
• Use retry loops with backoff for network calls and persist until success when needed.
• Use channels and defer unsubscribe/cleanup for deterministic resource release.

Example use cases

• Implementing agent startup sequence that posts machine info to backend and retries until success.
• Building a job runner that pulls tasks, executes them concurrently up to ParallelTaskCount, and reports results.
• Adding a lightweight collector that periodically gathers host metrics and posts them via the HTTP client.
• Implementing an in-place agent upgrade flow coordinated by the daemon process.
• Creating utility APIs for standardized HTTP calls that convert responses into DevopsResult objects.