Quick Start

SWEny is a platform for building AI-powered engineering workflows that follow a Learn → Act → Report pattern. SWEny Triage is the first recipe — it monitors your observability logs, investigates issues, creates tickets, and opens fix PRs. This guide walks through setting it up.

Prerequisites

• A GitHub repository
• An observability platform — Datadog, Sentry, CloudWatch, Splunk, Elasticsearch, New Relic, or Grafana Loki
• A Claude Max subscription (recommended) or an Anthropic API key

That’s all you need to get started. SWEny uses GitHub Issues by default — no extra issue tracker setup required.

Step 1: Try it (dry run)

Create .github/workflows/sweny-triage.yml in your repository:

name: SWEny Triage
on:
  workflow_dispatch:  # Manual trigger

permissions:
  contents: read
  issues: write

jobs:
  triage:
    runs-on: ubuntu-latest
    timeout-minutes: 60
    steps:
      - uses: actions/checkout@v4
        with:
          fetch-depth: 0

      - uses: swenyai/sweny@v0.2
        with:
          claude-oauth-token: ${{ secrets.CLAUDE_CODE_OAUTH_TOKEN }}
          dd-api-key: ${{ secrets.DD_API_KEY }}
          dd-app-key: ${{ secrets.DD_APP_KEY }}
          dry-run: true

Add secrets

In your repository Settings → Secrets and variables → Actions, add:

Secret	Where to get it
CLAUDE_CODE_OAUTH_TOKEN	Claude Max subscription → OAuth token
DD_API_KEY	Datadog → Organization Settings → API Keys
DD_APP_KEY	Datadog → Organization Settings → Application Keys

Why Claude Max? Predictable monthly cost — no per-token billing surprises. If you prefer pay-per-use, swap claude-oauth-token for anthropic-api-key with an Anthropic API key.

Run it

Go to the Actions tab, select “SWEny Triage”, and click Run workflow.

SWEny will analyze your Datadog logs and post a full investigation report in the GitHub Actions summary — no tickets, no PRs, just analysis. Review the output to see what it finds.

Step 2: Turn it on

Once you’re happy with the investigation quality, remove dry-run, add write permissions, and put it on a schedule:

name: SWEny Triage
on:
  schedule:
    - cron: '0 6 * * 1,4'  # Mon & Thu at 6 AM UTC
  workflow_dispatch:

permissions:
  contents: write
  pull-requests: write
  issues: write

jobs:
  triage:
    runs-on: ubuntu-latest
    timeout-minutes: 60
    steps:
      - uses: actions/checkout@v4
        with:
          fetch-depth: 0

      - uses: swenyai/sweny@v0.2
        with:
          claude-oauth-token: ${{ secrets.CLAUDE_CODE_OAUTH_TOKEN }}
          dd-api-key: ${{ secrets.DD_API_KEY }}
          dd-app-key: ${{ secrets.DD_APP_KEY }}

SWEny will now automatically investigate errors, create GitHub Issues with root cause analysis, write fixes, and open PRs.

Using a different observability provider?

Swap out the Datadog inputs for your provider. See Observability Providers for Sentry, CloudWatch, Splunk, Elasticsearch, New Relic, and Loki configuration.

Want Linear or Jira instead of GitHub Issues?

Set issue-tracker-provider and add the relevant credentials:

# Linear
- uses: swenyai/sweny@v0.2
  with:
    claude-oauth-token: ${{ secrets.CLAUDE_CODE_OAUTH_TOKEN }}
    dd-api-key: ${{ secrets.DD_API_KEY }}
    dd-app-key: ${{ secrets.DD_APP_KEY }}
    issue-tracker-provider: 'linear'
    linear-api-key: ${{ secrets.LINEAR_API_KEY }}
    linear-team-id: ${{ vars.LINEAR_TEAM_ID }}

See Issue Tracking Providers for full Linear, Jira, and GitHub Issues configuration.

What’s next?

• Provider Architecture — understand how SWEny’s plugin system works
• Action Inputs — configure time ranges, severity filters, investigation depth
• Examples — service filtering, specific issues, cross-repo dispatch