SWEny

npm install -g @sweny-ai/core

SWEny uses Claude as its AI engine. You’ll need an Anthropic API key, OAuth token, or authenticated Claude Code instance.

sweny workflow create "scan our repo for security issues, check dependencies \
  for vulnerabilities, and create Linear tickets for anything critical"

GitHub Security Audit

┌─────────────────────────────────┐
│ ○ Scan Commits for Secrets      │
└───────────────┬─────────────────┘
                │
┌───────────────▼─────────────────┐
│ ○ Scan Deps for Vulnerabilities │
└───────────────┬─────────────────┘
                │
┌───────────────▼─────────────────┐
│ ○ Compile Security Report       │
└───────────────┬─────────────────┘
                │
┌───────────────▼─────────────────┐
│ ○ Create Linear Tickets         │
└─────────────────────────────────┘

Save to security_audit.yml? [Y/n/refine]

sweny workflow run .sweny/workflows/github_security_audit.yml

✓ [1/4] Scan Commits for Exposed Secrets   12s
  ↳ Searched 847 commits, 0 secrets found
✓ [2/4] Scan Dependencies for Vulns        28s
  ↳ 2 critical CVEs in transitive deps
⠹ [3/4] Compile Security Report            6s
  ↳ Analyzing findings…

That’s it. One sentence in, a full workflow out. No YAML to write, no boilerplate, no config.

Don’t like the result? Edit it the same way you created it:

sweny workflow edit security_audit.yml "add a quality gate that rejects \
  vague findings, and loop back to re-scan if rejected"

---

Built-in workflows

SWEny ships two production-ready workflows you can run immediately:

Triage

Monitors your observability platform for errors, investigates root causes, creates issues, writes fixes, and opens PRs — automatically. Learn more

Implement

Takes an existing issue, analyzes the codebase, writes a fix, opens a PR, and notifies your team. Trigger it manually or chain it from Triage. Learn more

sweny triage --dry-run          # investigate production errors
sweny implement ENG-123         # fix a tracked issue and open a PR

Or build your own from a single sentence with sweny workflow create.

E2E browser testing

Generate AI-driven end-to-end tests for any web app — no test scripts to write:

sweny e2e init     # wizard asks about your flows → generates workflow YAML
sweny e2e run      # AI agent drives a real browser to test your app

The wizard supports 7 flow types — registration, login, purchase, onboarding, upgrade, cancellation, and custom. Each generates a self-contained workflow with auto-generated test credentials and optional cleanup. Get started with E2E testing

Beyond DevOps

SWEny workflows aren’t limited to triage and fixes. Teams use them to replace brittle data pipelines, automate research, and build quality-gated extraction systems.

Data Pipelines

Turn unstructured documents into structured data with LLM extraction, automated quality gates, and change detection. Read the deep-dive

Security Reviews

Ingest vendor questionnaires, draft answers from your security posture docs, and validate nothing is overstated.

See all use cases Competitive monitoring, contract management, feedback synthesis, and more.

Run it on a schedule

Deploy any workflow to GitHub Actions with three secrets and a cron schedule. SWEny monitors your production errors and opens fix PRs while you sleep.

.github/workflows/sweny-triage.yml

- uses: swenyai/triage@v1
  with:
    claude-oauth-token: ${{ secrets.CLAUDE_CODE_OAUTH_TOKEN }}
    observability-provider: sentry
    sentry-auth-token: ${{ secrets.SENTRY_AUTH_TOKEN }}

Full Action guide — setup, scheduling, and service maps.

Use it from Claude Code

Install the Claude Code plugin and get slash commands, MCP tools, and a startup hook — without leaving your conversation:

/plugin marketplace add swenyai/sweny
/plugin install sweny@sweny-official

Then type /sweny:triage, /sweny:implement ENG-123, /sweny:e2e-run, or any of the 11 available commands. Claude can also autonomously delegate tasks to SWEny’s DAG executor via MCP tools.

Connect to your stack

SWEny connects to the tools you already use — GitHub, Linear, Sentry, Datadog, BetterStack, Slack, and more — through built-in skills and MCP servers. Set an API key and the skill is ready. Bring your own MCP servers to connect anything else.

Studio

Studio is a visual DAG editor and live execution monitor. Design workflows by dragging nodes, or watch running workflows execute node-by-node in real time. Pass --stream to any CLI run to pipe events into Studio.

Workflow Marketplace

marketplace.sweny.ai — a community catalog of ready-to-run SWEny workflows. Browse by category, copy a workflow into .sweny/workflows/, customize the steps, and run it. The CLI and Action are free and always will be.