Kelos - The Kubernetes-native framework for orchestrating autonomous AI coding agents.
Self-dogfooding via 24/7 agent tasks on Kelos itself; solid but unproven at scale.
Platform engineers and teams running large Kubernetes clusters with autonomous development pipelines
Temporal · Dagger · GitHub Actions
The idea is that you define your everyday development workflows as YAML and let them run continuously on Kubernetes. Agent tasks are Kubernetes CRDs, so things like:
- Watch for “bug” issues → auto-draft a fix PR
- Auto-review incoming pull requests
- Auto-triage new issues with labels and priority
- Periodically scan the codebase → propose improvements
- Test the project as a new user → surface rough edges
Kelos is using this on itself. The self-development pipeline picks up open issues, investigates them, opens or updates a PR, self-reviews, and retries if CI fails. When the agent gets stuck, it labels the issue needs-input and stops: https://github.com/kelos-dev/kelos/tree/main/self-developmen...
I still do the final review and manually merge PRs, but it’s been quite helpful.
https://github.com/kelos-dev/kelos
Happy to answer questions about the design or what’s broken.
Kubernetes operator for autonomous AI agents—orchestrates Claude Code, Gemini, Codex as isolated, event-driven jobs.
Docker sandbox for Claude Code's dangerous flag when Anthropic won't let you run it bare.
Auto-detects docker-compose.yml so agents can test against real database dependencies.
Actual Kubernetes operator for agent lifecycle, but orchestrating agents is still a niche use case.
OS-level sandboxing blocks base64 evasion when pattern matching alone fails.
Kubernetes-native AI agent orchestration via YAML; early but self-hosting Kelos development proves the loop.