{"id":48271126,"url":"https://github.com/kubeswarm/kubeswarm","last_synced_at":"2026-04-20T18:01:06.741Z","repository":{"id":350736723,"uuid":"1207988702","full_name":"kubeswarm/kubeswarm","owner":"kubeswarm","description":"Kubernetes operator that manages AI agents as first-class resources","archived":false,"fork":false,"pushed_at":"2026-04-18T21:36:07.000Z","size":877,"stargazers_count":1,"open_issues_count":1,"forks_count":0,"subscribers_count":0,"default_branch":"main","last_synced_at":"2026-04-18T23:25:05.227Z","etag":null,"topics":["a2a","agentic-workflow","agentic-workflows-orchestration","agents","ai-agents","golang","kubernetes","kubernetes-operator","llm","mcp"],"latest_commit_sha":null,"homepage":"https://docs.kubeswarm.io","language":"Go","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"apache-2.0","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/kubeswarm.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":"CONTRIBUTING.md","funding":null,"license":"LICENSE","code_of_conduct":"CODE_OF_CONDUCT.md","threat_model":null,"audit":null,"citation":null,"codeowners":".github/CODEOWNERS","security":"SECURITY.md","support":null,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null,"zenodo":null,"notice":null,"maintainers":null,"copyright":null,"agents":null,"dco":null,"cla":null}},"created_at":"2026-04-11T17:07:18.000Z","updated_at":"2026-04-18T21:19:13.000Z","dependencies_parsed_at":null,"dependency_job_id":null,"html_url":"https://github.com/kubeswarm/kubeswarm","commit_stats":null,"previous_names":["kubeswarm/kubeswarm"],"tags_count":2,"template":false,"template_full_name":null,"purl":"pkg:github/kubeswarm/kubeswarm","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/kubeswarm%2Fkubeswarm","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/kubeswarm%2Fkubeswarm/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/kubeswarm%2Fkubeswarm/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/kubeswarm%2Fkubeswarm/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/kubeswarm","download_url":"https://codeload.github.com/kubeswarm/kubeswarm/tar.gz/refs/heads/main","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/kubeswarm%2Fkubeswarm/sbom","scorecard":null,"host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":286080680,"owners_count":32059139,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2026-04-20T11:35:06.609Z","status":"ssl_error","status_checked_at":"2026-04-20T11:34:48.899Z","response_time":94,"last_error":"SSL_connect returned=1 errno=0 peeraddr=140.82.121.6:443 state=error: unexpected eof while reading","robots_txt_status":"success","robots_txt_updated_at":"2025-07-24T06:49:26.215Z","robots_txt_url":"https://github.com/robots.txt","online":false,"can_crawl_api":true,"host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":["a2a","agentic-workflow","agentic-workflows-orchestration","agents","ai-agents","golang","kubernetes","kubernetes-operator","llm","mcp"],"created_at":"2026-04-04T22:15:14.779Z","updated_at":"2026-04-20T18:01:06.729Z","avatar_url":"https://github.com/kubeswarm.png","language":"Go","funding_links":[],"categories":[],"sub_categories":[],"readme":"\u003cp align=\"center\"\u003e\n  \u003cimg src=\"https://assets.kubeswarm.io/logo.svg\" width=\"200\" alt=\"kubeswarm\"\u003e\n\u003c/p\u003e\n\n\u003ch3 align=\"center\"\u003eAgents are workloads. Manage them like it.\u003c/h3\u003e\n\n\u003cp align=\"center\"\u003e\n  \u003ca href=\"https://github.com/kubeswarm/kubeswarm/actions/workflows/ci.yml\"\u003e\u003cimg src=\"https://github.com/kubeswarm/kubeswarm/actions/workflows/ci.yml/badge.svg\" alt=\"CI\"\u003e\u003c/a\u003e\n  \u003ca href=\"https://github.com/kubeswarm/kubeswarm/releases\"\u003e\u003cimg src=\"https://img.shields.io/github/v/release/kubeswarm/kubeswarm?include_prereleases\" alt=\"GitHub release\"\u003e\u003c/a\u003e\n  \u003ca href=\"./LICENSE\"\u003e\u003cimg src=\"https://img.shields.io/badge/license-Apache%202.0-blue.svg\" alt=\"License\"\u003e\u003c/a\u003e\n  \u003ca href=\"go.mod\"\u003e\u003cimg src=\"https://img.shields.io/github/go-mod/go-version/kubeswarm/kubeswarm\" alt=\"Go version\"\u003e\u003c/a\u003e\n  \u003ca href=\"https://goreportcard.com/report/github.com/kubeswarm/kubeswarm\"\u003e\u003cimg src=\"https://goreportcard.com/badge/github.com/kubeswarm/kubeswarm\" alt=\"Go Report Card\"\u003e\u003c/a\u003e\n\u003c/p\u003e\n\nKubernetes operator that manages AI agents as first-class resources. Define agents in YAML, connect MCP tools, compose multi-agent pipelines, and operate with the same tooling you already use for services.\n\n\u003e **Status: v0.2.0-alpha** - Core primitives are functional. API is `v1alpha1` and may change between minor versions. Not recommended for production workloads yet. See [VERSIONING.md](./VERSIONING.md).\n\n---\n\n## What it does\n\nkubeswarm introduces CRDs that model every part of an AI agent deployment:\n\n| CRD               | Purpose                                                                     |\n| ----------------- | --------------------------------------------------------------------------- |\n| **SwarmAgent**    | LLM agent with model, system prompt, MCP tools, guardrails, and autoscaling |\n| **SwarmTeam**     | Multi-agent pipeline (DAG, sequential, or LLM-routed)                       |\n| **SwarmRun**      | Single execution of a team pipeline with audit trail                        |\n| **SwarmBudget**   | Token spend tracking and enforcement per team/namespace                     |\n| **SwarmRegistry** | Agent capability discovery and delegation                                   |\n| **SwarmSettings** | Shared configuration (MCP servers, context policies)                        |\n| **SwarmMemory**   | Vector memory backends (pgvector, Qdrant) for agent recall                  |\n| **SwarmEvent**    | Trigger pipeline runs from external events (webhooks, cron)                 |\n| **SwarmNotify**   | Run completion notifications (webhook, Slack)                               |\n| **SwarmPolicy**   | Governance constraints (model allow/deny, token limits, tool restrictions)  |\n\nAll resources are namespace-scoped. `kubectl get kubeswarm -A` shows everything.\n\n## Key features\n\n- **Multi-provider** - Anthropic, OpenAI, Google Gemini, or any provider via gRPC plugin\n- **Reasoning support** - Anthropic extended thinking, OpenAI reasoning effort, guardrail clamping\n- **MCP tool integration** - connect any MCP server; dynamic tool discovery; per-tool trust levels\n- **Agent-to-agent** - gateway dispatch, advisor consultations, and capability-based routing across agents\n- **Pipeline orchestration** - DAG-based, sequential, or LLM-routed dispatch with step validation\n- **Governance** - SwarmPolicy enforces model allow/deny lists, token limits, and tool restrictions across namespaces\n- **Cost controls** - per-agent token limits, daily budgets, circuit breakers, spend tracking\n- **Observability** - OTel metrics and traces, structured audit trail, MCP health monitoring\n- **Security** - pod hardening, network policies, prompt injection defense, tool allow/deny lists\n- **Autoscaling** - KEDA-based scale-to-zero and demand-driven replica management\n\n---\n\nFull documentation at **[docs.kubeswarm.io](https://docs.kubeswarm.io)**.\n\n---\n\n## Prerequisites\n\n- Kubernetes 1.35+\n- kubectl\n- Docker 24+\n- kind 0.27+ (for local development)\n\n## Quick start (local development)\n\n```bash\n# 1. Clone and setup\ngit clone https://github.com/kubeswarm/kubeswarm.git\ncd kubeswarm\nmake setup\n\n# 2. Run the full CI pipeline locally\nmake ci\n\n# 3. Deploy to a local Kind cluster\nmake local-up\n\n# 4. Verify the controller is running\nkubectl get pods -n kubeswarm-system\n```\n\n## Quick start (Helm)\n\n```bash\n# 1. Add the Helm repo\nhelm repo add kubeswarm https://kubeswarm.github.io/helm-charts/\nhelm repo update\n\n# 2. Install the operator\nhelm install kubeswarm kubeswarm/kubeswarm \\\n  --namespace kubeswarm-system --create-namespace\n\n# 3. Create a Secret with your LLM API key\nkubectl create secret generic llm-api-key \\\n  --namespace default \\\n  --from-literal=ANTHROPIC_API_KEY=sk-ant-...\n\n# 4. Apply a sample team\nkubectl apply -f https://raw.githubusercontent.com/kubeswarm/kubeswarm-cookbook/main/teams/01-simple-pipeline/blog-writer.yaml\n\n# 5. Trigger a run\nswarm trigger blog-writer-team -n default \\\n  --input '{\"topic\": \"Kubernetes operators explained\"}'\n\n# 6. Watch it run\nswarm status -n default\n```\n\nInstall the `swarm` CLI: [kubeswarm-cli](https://github.com/kubeswarm/kubeswarm-cli).\n\n## Project structure\n\n```\nkubeswarm/\n  api/v1alpha1/          CRD type definitions\n  internal/controller/   Reconcilers (one per CRD)\n  internal/webhook/      Admission webhooks\n  internal/mcpgateway/   MCP SSE gateway for agent-to-agent calls\n  pkg/                   Shared packages (audit, costs, healthz, observability)\n  runtime/               Nested Go module - agent binary + vendor SDK implementations\n    cmd/kubeswarm-runtime/      Agent runtime entrypoint\n    cmd/kubeswarm-controller/   Controller binary entrypoint\n    pkg/providers/       LLM provider implementations (Anthropic, OpenAI, Gemini)\n    pkg/queue/           Task queue backends (Redis)\n    pkg/vectors/         Vector store backends (pgvector, Qdrant)\n    pkg/artifacts/       Artifact store backends (S3, GCS)\n```\n\nThe core `kubeswarm/` module has zero vendor SDK imports. All LLM, queue, and storage\nSDK dependencies live in `runtime/` (nested module) to keep the operator's dependency\ntree clean.\n\n## Related repos\n\n| Repo                                                                  | Description                                  |\n| --------------------------------------------------------------------- | -------------------------------------------- |\n| [helm-charts](https://github.com/kubeswarm/helm-charts)               | Helm chart for operator deployment           |\n| [kubeswarm-cli](https://github.com/kubeswarm/kubeswarm-cli)           | Local dev CLI (`swarm run`, `swarm trigger`) |\n| [kubeswarm-docs](https://github.com/kubeswarm/kubeswarm-docs)         | Documentation site (docs.kubeswarm.io)       |\n| [kubeswarm-cookbook](https://github.com/kubeswarm/kubeswarm-cookbook) | Example pipelines and recipes                |\n\n## Contributing\n\nIssues, ideas and PRs are welcome. See [CONTRIBUTING.md](./CONTRIBUTING.md) for guidelines.\n\n## License\n\nApache 2.0 - see [LICENSE](./LICENSE)\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fkubeswarm%2Fkubeswarm","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fkubeswarm%2Fkubeswarm","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fkubeswarm%2Fkubeswarm/lists"}