We build polyglot, .NET-first software with an AI-assisted development workflow. Our public repositories are currently the engineering platform behind that work: shared quality standards, CI building blocks, isolated runners, and the agent tooling we develop with.
- standards: shared repository-quality configs and engineering/review conventions, distributed to consuming repos and read by people and agents alike.
- ci-workflows: SHA-pinned, configurable composite actions (one per code-quality tool) and reusable GitHub Actions workflows, aggregated by consumers into a single status gate.
- ci-runner: an official Actions runner image plus a native Windows scale-set controller. Each job gets a fresh, one-job Docker worker, with GitHub-hosted runners as fallback.
- Reusable, repo-agnostic agent tooling, covering skills, agents, hooks, and MCP servers, published as one marketplace per assistant: claude-code-plugins, codex-plugins, and cursor-plugins.
- The
standardsrepo states the quality bar as configs and conventions that people and agents both read. - Repositories enforce that bar by calling the same SHA-pinned
ci-workflowsactions. - GitHub governance is managed as infrastructure-as-code so settings are reviewable and reproducible.
- AI assists implementation and review; automated checks and maintainer judgment remain the release gates.
- Browse the repositories above, and open an issue or pull request where you see something to improve.
- Review our Code of Conduct and contribution guidelines before contributing.
- Found a security issue? Follow our security policy and do not report vulnerabilities publicly.