setup-ci

name	setup-ci
description	Configure opencode-actions GitHub Actions workflows for a repository. Use when the user wants to set up automated PR review, multi-review, or comment-triggered opencode commands in their CI. Triggers on requests like "帮我配置 CI", "设置 PR review", "添加 opencode action", "配置自动审查", or any task involving adding opencode-actions workflows to a GitHub repository.

Setup opencode-actions CI

Configure sun-praise/opencode-actions GitHub Actions for a user's repository.

Workflow

1. Ask the user which actions they want and their repository info
2. Check the action reference for available options: see references/actions-reference.md
3. Generate the workflow YAML file(s)
4. Guide the user on secrets and permissions setup

Action Selection

Need	Action	One-liner
Multi-agent review	multi-review	Parallel reviewer personas + coordinator synthesis
Custom command	github-run-opencode	Flexible, user-defined prompt
Manual setup	setup-opencode + run-opencode	Full control over install and run

multi-review includes 8 built-in reviewer personas: quality, security, performance, architecture, regression-test, feature-missing, test-value, spec-coverage. Use default-team to select which personas run.

Recommended Models

Model	Provider	Required Secret	Notes
deepseek/deepseek-v4-flash	DeepSeek	DEEPSEEK_API_KEY	Fast and cost-effective; recommended for multi-review
zhipuai-coding-plan/glm-5.1	Zhipu	ZHIPU_API_KEY	General-purpose, good balance of speed and quality
opencode-go/deepseek-v4-flash	OpenCode Go	OPENCODE_GO_API_KEY	Proxy service, uses DeepSeek under the hood
minimax-cn-coding-plan/MiniMax-M3	MiniMax	MINIMAX_API_KEY	Chinese-language review; not compatible with multi-review (why?)
xiaomi-token-plan-cn/mimo-v2-pro	Xiaomi MiMo	XIAOMI_API_KEY	Token Plan (China); not compatible with multi-review (why?)

Set via model: input in the with: block (e.g. model: ${{ vars.MODEL_NAME }}), or configure MODEL_NAME as a repository variable in Settings → Secrets and variables → Actions → Variables to switch models without modifying workflow files.

Multi-Review Setup

Generate this in .github/workflows/opencode-multi-review.yml:

name: OpenCode Multi-Review

on:
  pull_request:
    types: [opened, synchronize, reopened, ready_for_review]

jobs:
  multi-review:
    if: github.event.pull_request.draft == false && github.event.pull_request.head.repo.full_name == github.repository
    runs-on: ubuntu-latest
    permissions:
      contents: read
      pull-requests: write
      issues: write
    steps:
      - name: Checkout PR head
        uses: actions/checkout@v6
        with:
          repository: ${{ github.event.pull_request.head.repo.full_name }}
          ref: ${{ github.event.pull_request.head.ref }}

      - name: Run multi-review
        uses: sun-praise/opencode-actions/multi-review@v4
        with:
          github-token: ${{ secrets.GITHUB_TOKEN }}
          deepseek-api-key: ${{ secrets.DEEPSEEK_API_KEY }}
          zhipu-api-key: ${{ secrets.ZHIPU_API_KEY }}
          opencode-go-api-key: ${{ secrets.OPENCODE_GO_API_KEY }}
          litellm-url: ${{ secrets.LITELLM_URL }}
          litellm-api-key: ${{ secrets.LITELLM_API_KEY }}
          # Optional: override default reviewer team
          # default-team: "quality:2,security:1,architecture:1,test-value:1"
          # Optional: exclude files from diff (e.g. "*.lock,*.snap")
          # diff-exclude: "*.lock,*.snap"
          # Optional: truncate diff over N KB (default: 0 = disabled)
          # diff-max-size-kb: "200"
          # Optional: response language: "zh" (default) or "en"
          # language: "zh"
          # Optional: increase timeout for large PRs (default: 900s)
          # timeout-seconds: "1200"

Minimal Review Setup (single persona)

Generate this in .github/workflows/opencode-review.yml:

name: OpenCode PR Review

on:
  pull_request:
    types: [opened, synchronize, reopened, ready_for_review]

jobs:
  review:
    if: github.event.pull_request.draft == false && github.event.pull_request.head.repo.full_name == github.repository
    runs-on: ubuntu-latest
    permissions:
      contents: read
      pull-requests: write
      issues: write
    steps:
      - name: Checkout PR head
        uses: actions/checkout@v6
        with:
          repository: ${{ github.event.pull_request.head.repo.full_name }}
          ref: ${{ github.event.pull_request.head.ref }}

      - name: Run OpenCode review
        uses: sun-praise/opencode-actions/multi-review@v4
        with:
          github-token: ${{ secrets.GITHUB_TOKEN }}
          zhipu-api-key: ${{ secrets.ZHIPU_API_KEY }}
          opencode-go-api-key: ${{ secrets.OPENCODE_GO_API_KEY }}
          deepseek-api-key: ${{ secrets.DEEPSEEK_API_KEY }}
          litellm-url: ${{ secrets.LITELLM_URL }}
          litellm-api-key: ${{ secrets.LITELLM_API_KEY }}
          default-team: "quality:1"

Comment Command Setup

Generate this in .github/workflows/opencode-command.yml:

name: OpenCode Comment Command

on:
  issue_comment:
    types: [created]
  pull_request_review_comment:
    types: [created]

jobs:
  opencode:
    if: >-
      (contains(github.event.comment.body, '/oc') || contains(github.event.comment.body, '/opencode')) &&
      (github.event.comment.author_association == 'OWNER' ||
       github.event.comment.author_association == 'MEMBER' ||
       github.event.comment.author_association == 'COLLABORATOR')
    runs-on: ubuntu-latest
    permissions:
      id-token: write
      contents: write
      pull-requests: write
      issues: write
    steps:
      - name: Resolve checkout target
        id: target
        uses: actions/github-script@v8
        with:
          script: |
            const core = require('@actions/core')
            let repository = `${context.repo.owner}/${context.repo.repo}`
            let ref = context.sha
            let isFork = false

            if (context.eventName === 'pull_request_review_comment') {
              repository = context.payload.pull_request.head.repo.full_name
              ref = context.payload.pull_request.head.sha
              isFork = context.payload.pull_request.head.repo.fork
            } else if (context.eventName === 'issue_comment' && context.payload.issue.pull_request) {
              const pull = await github.rest.pulls.get({
                owner: context.repo.owner,
                repo: context.repo.repo,
                pull_number: context.payload.issue.number,
              })
              repository = pull.data.head.repo.full_name
              ref = pull.data.head.sha
              isFork = pull.data.head.repo.fork
            }

            core.setOutput('repository', repository)
            core.setOutput('ref', ref)
            core.setOutput('is_fork', isFork ? 'true' : 'false')

      - name: Skip fork PR commands
        if: ${{ steps.target.outputs.is_fork == 'true' }}
        shell: bash
        run: |
          set -euo pipefail
          printf 'Skipping OpenCode command on forked pull request comments to avoid exposing repository secrets.\n'

      - name: Checkout repository
        if: ${{ steps.target.outputs.is_fork != 'true' }}
        uses: actions/checkout@v6
        with:
          repository: ${{ steps.target.outputs.repository }}
          ref: ${{ steps.target.outputs.ref }}

      - name: Run OpenCode
        if: ${{ steps.target.outputs.is_fork != 'true' }}
        uses: sun-praise/opencode-actions/github-run-opencode@v4
        env:
          MODEL_NAME: zhipuai-coding-plan/glm-5.1
        with:
          github-token: ${{ secrets.GITHUB_TOKEN }}
          zhipu-api-key: ${{ secrets.ZHIPU_API_KEY }}
          opencode-go-api-key: ${{ secrets.OPENCODE_GO_API_KEY }}
          litellm-url: ${{ secrets.LITELLM_URL }}
          litellm-api-key: ${{ secrets.LITELLM_API_KEY }}

Custom Reviewer Personas

Users can add custom reviewer personas by placing .yaml or .yml files in .github/reviewers/:

# .github/reviewers/accessibility.yaml
name: accessibility
prompt: |
  Review this PR for accessibility issues...

Custom personas are referenced in default-team just like built-in ones: "accessibility:1,quality:1".

Available Personas

Persona	Focus
quality	Code quality, bugs, style
security	Security vulnerabilities, injection, auth
performance	Performance issues, hot paths, resource leaks
architecture	Coupling, layering, module placement, structural concerns
regression-test	Missing regression tests for bug fixes
feature-missing	Missing features vs linked issue spec
test-value	Low-value tests: empty assertions, hardcoded mocks, duplicates
spec-coverage	Missing features vs project spec/task files

Secrets Setup

The user needs to configure at least one provider API key in their repository secrets:

1. Go to the repository on GitHub
2. Settings → Secrets and variables → Actions → New repository secret
3. Add one or more:
   • DEEPSEEK_API_KEY (recommended for multi-review)
   • ZHIPU_API_KEY
   • OPENCODE_GO_API_KEY
   • LITELLM_URL + LITELLM_API_KEY
   • MINIMAX_API_KEY (not compatible with multi-review)
   • XIAOMI_API_KEY (not compatible with multi-review)

GITHUB_TOKEN is automatically available in GitHub Actions.

Version Pinning

• Use @v4 for compatible updates within v4
• Security-sensitive repos should pin to a full commit SHA instead of @v4
• Immutable releases use semver tags like v4.0.0

For the complete list of inputs per action, see references/actions-reference.md.