automate-github-issues

작성자: google-labs-code

병렬 Jules 코딩 에이전트를 사용하여 자동화된 GitHub 이슈 분류 및 해결 설정

npx skills add https://github.com/google-labs-code/jules-skills --skill automate-github-issues

Automate GitHub Issues with Jules

You are setting up a repository to automatically analyze open GitHub issues, plan implementation tasks, and dispatch parallel Jules coding agents to fix them.

What You're Setting Up

A 5-phase automated pipeline that runs via GitHub Actions (or locally):

  1. Analyze — Fetch open issues and format as structured markdown
  2. Plan — A Jules session performs deep code-level triage and produces self-contained task prompts
  3. Validate — Verify no two tasks modify the same file (prevents merge conflicts)
  4. Dispatch — Spawn parallel Jules sessions, one per task
  5. Merge — Sequential PR merge with CI validation

Setup Steps

Step 1: Copy fleet scripts to the repository

Copy the entire scripts/ directory from this skill into the target repository at scripts/fleet/:

Target structure:
scripts/fleet/
├── fleet-analyze.ts
├── fleet-plan.ts
├── fleet-dispatch.ts
├── fleet-merge.ts
├── types.ts
├── prompts/
│   ├── analyze-issues.ts
│   └── bootstrap.ts
└── github/
    ├── git.ts
    ├── issues.ts
    ├── markdown.ts
    └── cache-plugin.ts

Important: Preserve the directory structure exactly. The scripts use relative imports between files.

Step 2: Copy workflow templates

Copy the workflow files from assets/ to the repository's .github/workflows/ directory:

  • assets/fleet-dispatch.yml.github/workflows/fleet-dispatch.yml
  • assets/fleet-merge.yml.github/workflows/fleet-merge.yml

Step 3: Create a package.json for the fleet scripts

Create scripts/fleet/package.json with the required dependencies:

{
  "name": "fleet-scripts",
  "private": true,
  "type": "module",
  "dependencies": {
    "@google/jules-sdk": "^0.1.0",
    "octokit": "^4.1.0",
    "find-up": "^7.0.0"
  },
  "devDependencies": {
    "@types/bun": "^1.2.0"
  }
}

Step 4: Create environment template

Copy assets/.env.example to the repository root.

Step 5: Install dependencies

cd scripts/fleet && bun install

Step 6: Print next steps for the user

Tell the user they need to:

  1. Add JULES_API_KEY as a GitHub repository secret (Settings → Secrets → Actions)
  2. GITHUB_TOKEN is provided automatically by GitHub Actions
  3. Customize the cron schedule in .github/workflows/fleet-dispatch.yml (default: daily 6am UTC)
  4. Commit all generated files

Manual Usage

After setup, the user can run the pipeline locally:

cd scripts/fleet

# Fetch open issues
bun fleet-analyze.ts

# Plan tasks (creates a Jules planning session)
JULES_API_KEY=<key> bun fleet-plan.ts

# Dispatch parallel agents
JULES_API_KEY=<key> bun fleet-dispatch.ts

# Merge PRs sequentially
GITHUB_TOKEN=<token> bun fleet-merge.ts

Customization

Prompt Tuning

The analysis prompt in scripts/fleet/prompts/analyze-issues.ts controls how deeply issues are investigated. Users can adjust:

  • Root cause analysis depth
  • Solution implementation detail level
  • Merge conflict avoidance rules
  • File ownership constraints

Issue Filtering

Edit scripts/fleet/github/issues.ts to filter issues by label, milestone, or state.

Resource References

Troubleshooting

  • "Unable to parse git remote URL": Ensure the repo has a valid GitHub remote (git remote get-url origin)
  • Ownership conflict errors: Two tasks claim the same file. Adjust the task JSON or merge them manually.
  • CI timeout during merge: Increase maxWaitMs in fleet-merge.ts (default: 10 minutes)
  • Bun not found: Install Bun: curl -fsSL https://bun.sh/install | bash

google-labs-code의 다른 스킬

agent-dx-cli-scale
google-labs-code
CLI가 AI 에이전트에 얼마나 잘 설계되었는지 평가하기 위한 점수 척도로, "AI 에이전트를 위한 CLI 재작성" 원칙에 기반합니다.
official
ink
google-labs-code
Ink 터미널 렌더러로, JSON 사양을 대화형 터미널 UI로 변환합니다. @json-render/ink로 작업하거나 터미널 UI를 구축할 때 사용하세요.
official
tdd-red-green-refactor
google-labs-code
이 스킬은 AI 지원 프로그래밍을 위한 구조적 프레임워크를 구현하여 모든 코드 라인이 검증 가능하고, 타입이 지정되며, 목적에 부합하도록 보장합니다.
official
typed-service-contracts
google-labs-code
Spec and Handler" 패턴을 사용하여 견고하고 타입 안전한 TypeScript 서비스를 구축하기 위한 아키텍처 표준입니다. CLI, 라이브러리 또는 복잡한...
official
local-action-verification
google-labs-code
act를 사용하여 로컬 GitHub Actions 검증을 위한 리포지토리를 설정하여, Jules가 푸시 전에 CI를 검증할 수 있도록 합니다.
official
github-codebase-briefing
google-labs-code
이 스킬은 저장소에 대한 "심층 읽기"를 수행합니다. 프로젝트의 아키텍처를 먼저 이해한 후, 해당 맥락에 맞춰 열린 항목들을 평가함으로써 피상적인 목록화를 피합니다.
official
stitch-sdk-bug-bash
google-labs-code
실제 API 키를 사용하여 Stitch SDK의 버그를 찾습니다. 표준 기능 경계와 까다로운 상황을 다룹니다.
official
stitch-sdk-development
google-labs-code
Stitch SDK를 개발합니다. 생성 파이프라인, 이중 모드(에이전트 대 SDK), 오류 처리, 트래픽 라이트(빨간색-초록색-노란색) 구현을 다룹니다…
official