smithery.ai

gh-workflow

Trigger and monitor GitHub Actions workflows, CI/CD pipelines, and automated builds. Use when running workflows, triggering CI, deploying via GitHub Actions, downloading artifacts, or checking workflow status on github.com

First seen Mar 28, 2026

Installation

$ npx skills add https://smithery.ai

Similar popular skills

Related neighbors and high-traction skills in the same topics — useful to compare before installing.

Also in this package

Other skills from smithery.ai · top by installs.

npx skills add https://smithery.ai

Browse all from smithery.ai

More details

Agent compatibility

Declared targets from SKILL.md / docs. Unmarked agents are not listed — the skill may still install via the CLI.

Claude Code Not declared
Cursor Not declared
Codex Not declared
GitHub Copilot Not declared
Windsurf Not declared
Gemini CLI Not declared
Cline Not declared
OpenCode Not declared

Package contents

Files included with this skill beyond the listing page.

  • skill md SKILL.md 2,756 B
  • docs SUMMARY.md 241 B

History

  1. First seen on skills.sh
  2. First recorded snapshot · 1 installs

SKILL.md

gh-workflow

Run GitHub workflows on github.com

Workflow

When the user asks to run a workflow:

  1. Verify git remotes and confirm commits are pushed (MANDATORY):

- Check remotes: git remote -v | grep github.com - Get current commit: git rev-parse HEAD - Verify commit exists on remote: git fetch github && git branch -r --contains $(git rev-parse HEAD) | grep -q github/ - If not found, push first: git push github HEAD:main (or appropriate branch name) - Quick check: git fetch github && git branch -r --contains HEAD | grep -q github/ || echo "WARNING: HEAD not pushed to github"

  1. Check the workflow file in .github/workflows/ to understand its trigger type and inputs
  2. For workflow_dispatch workflows: trigger directly with gh workflow run
  3. For push workflows: push to the remote first, then the workflow runs automatically
  4. Use appropriate parameters based on the workflow's input definitions
  5. For testing/iteration with matrix builds: prefer x64 over arm64 (faster provisioning)
  6. Download workflow artifacts to verify outputs

Usage

# List workflows
GH_HOST=github.com gh workflow list

# Run a workflow
GH_HOST=github.com gh workflow run <workflow-name>

# Run with inputs
GH_HOST=github.com gh workflow run <workflow-name> -f key=value

# View recent runs
GH_HOST=github.com gh run list --workflow=<workflow-name>

# Watch a run
GH_HOST=github.com gh run watch <run-id>

# Download artifacts from a run
GH_HOST=github.com gh run download <run-id>

Testing matrix builds

For workflows with matrix builds, test with the fastest variant first:

  • Prefer x64 over arm64 (x64 runners provision faster)
  • Prefer ubuntu over macos (faster startup)
  • Use workflow inputs to target specific matrix cells if supported

Examples

# Run workflow_dispatch workflow with inputs
GH_HOST=github.com gh workflow run luajit.yml -f release_tag=2025.11.23 -f create_release=true

# Run workflow at specific commit (preferred - explicit and reproducible)
GH_HOST=github.com gh workflow run luajit.yml --ref $(git rev-parse HEAD)

# Or with explicit SHA
GH_HOST=github.com gh workflow run luajit.yml --ref abc1234

# Run without creating release (testing)
GH_HOST=github.com gh workflow run nvim.yml -f create_release=false

# Download artifacts from latest run
GH_HOST=github.com gh run list --workflow=luajit.yml --limit 1 --json databaseId --jq '.[0].databaseId' | xargs -I {} GH_HOST=github.com gh run download {}