ci-workflow-guide

Design and maintain CI pipelines for SGLang projects with stage ordering and gating.

Updated Jun 29, 2026
One-click install
npx skills add https://github.com/hey-kong/sglang --skill ci-workflow-guide-hey-kong
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: ci-workflow-guide
Source: https://github.com/hey-kong/sglang/tree/main/.claude/skills/ci-workflow-guide
Command: npx skills add https://github.com/hey-kong/sglang --skill ci-workflow-guide-hey-kong

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill provides a comprehensive blueprint for building and maintaining robust CI pipelines for SGLang projects, including stage ordering, fast-fail, gating, partitioning, and debugging CI failures across stages.

Core Features & Use Cases

  • Stage ordering and parallelism: orchestrates A → B → C or parallel stage execution to optimize feedback loops.
  • Gating and fast-fail: implements health checks, branch gating, per-stage failure isolation, and root-cause annotation to accelerate debugging.
  • Partitioning and scalability: uses LPT-based test partitioning and matrix strategies to balance workloads across runners.
  • Use Case: teams modifying CI workflows can quickly understand test registration, wait-for-stage logic, and how to add new stages with consistent gating.

Quick Start

Define or adjust your PR workflow to enable gating and wait-for-stage with health checks using the guidance in this Skill.

Frequently Asked Questions about ci-workflow-guide

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I implement fast-fail and gating in GitHub Actions CI pipelines?

Add new CI stages by following the structural and behavioral requirements for consistent gating. You need to configure test registration, implement wait-for-stage logic, and apply cross-stage synchronization to maintain pipeline orchestration.

What is the best way to partition tests across CI runners for workload balancing?

The best way to partition tests across CI runners is using LPT-based test partitioning combined with matrix strategies. This balances workloads across runners, optimizing feedback loops and ensuring cross-stage synchronization during parallel execution.

How does wait-for-stage logic work in SGLang CI workflow orchestration?

Wait-for-stage logic in SGLang CI workflow orchestration works by applying cross-stage synchronization and health checks before proceeding. It ensures dependent stages wait for prior stage completion, providing clear root-cause reporting for any detected failures.

Can I use this CI workflow orchestration for scheduled and manual runs instead of just PR-gating?

Yes, you can use this CI workflow orchestration for scheduled and manual runs instead of just PR-gating. The blueprint applies to PR-gated, scheduled, and manual runs, detailing stage ordering, fast-fail, partitioning, and debugging across all execution triggers.