Building practical open source tools for AI workflows, developer systems, and automation that ships.
Dallay is an open source lab for tools that improve how modern teams build software.
This org is intentionally not locked into one product category. Some projects become developer tools, some become infrastructure building blocks, and some grow into bigger platform bets. The through line is consistent: useful software, clean interfaces, portable workflows, and systems that hold up outside the demo environment.
| Project | What it is | Why it matters |
|---|---|---|
corvus |
An always-on AI agent platform for long-running orchestration workloads. | This is the bigger platform direction: autonomous systems that can run with structure, context, and operational realism. |
agentsync |
A CLI for synchronizing AI agent instructions, commands, skills, and MCP config across coding assistants. | It solves a real ecosystem problem: too many tools, too many config formats, too much duplication. |
common-actions |
Shared GitHub Actions and reusable workflows for the Dallay ecosystem. | It keeps CI/CD consistent, reduces copy-paste automation, and makes repos easier to scale. |
- AI tooling with a strong bias toward real usage, not hype.
- CLI utilities that remove setup friction and repeated manual work.
- Reusable automation for CI/CD, releases, validation, and repo hygiene.
- Internal platform pieces that later become good open source citizens.
- Build composable tools that work well alone and better together.
- Prefer reliability, clarity, and maintainability over cleverness.
- Keep humans in the loop where trust, review, or judgment matters.
- Document decisions clearly so adoption is easier than reverse engineering.
- Improve incrementally, release often, and let real usage shape the roadmap.
- Explore
corvusif you care about agent platforms and orchestration. - Try
agentsyncif you use multiple AI coding assistants and want one source of truth. - Adopt
common-actionsif you want cleaner, reusable GitHub automation.
If something here helps you, you can support the work in a few simple ways:
- Open issues when you hit bugs, sharp edges, or missing docs.
- Start discussions when you see a pattern worth turning into a reusable tool.
- Send pull requests for fixes, docs improvements, tests, or features.
- Star the repos you want to see pushed further.
For contribution details, check each repository's own guidelines and conventions.
More projects will show up here as they move from experiments to tools worth maintaining in public.