SWEny, YAML workflows for AI agents I'm running in prod (triage, E2E)

What's New

Providers

• 7 new providers: Splunk, Elasticsearch, Grafana Loki, New Relic, GitLab, OpsGenie, Microsoft Teams
• AWS Secrets Manager credential vault provider
• Kubernetes CSI storage provider
• Jira issue tracking provider
• ProviderApiError — all 18+ providers now throw structured errors with provider name, HTTP status, and response body
• Zod schema validation for Slack messaging config (last provider without it)

Engine

• Configurable base branch — repos using master or develop no longer fail (baseBranch config)
• Configurable PR labels — override default ["agent", "triage", "needs-review"]
• Configurable analysis directory — analysisDir replaces hardcoded .github/triage-analysis/
• Configurable PR description turns and issue priority
• Peer dependency on @swenyai/providers declared
• Dependency constraint tightened from "*" to "^0.2.0"
• 34 new prompt builder tests covering all 3 prompt functions
• Unsafe type cast fixes in investigate.ts and elastic.ts

Action

• Input validation — fail-fast with clear messages when credentials are missing (auth, observability, issue tracker, source control, notification)
• base-branch and pr-labels inputs added to action.yml
• 16 new validation tests

Infrastructure

• Husky + lint-staged pre-commit hooks
• JSDoc documentation on all provider interfaces
• ESLint warnings reduced from 173 to 18
• 660+ tests across all packages

npm Packages

npm install @wickdninja/[email protected]
npm install @wickdninja/[email protected]

Full Changelog: https://github.com/swenyai/sweny/compare/v0.1...v0.2.0