Open source infrastructure for running, connecting, and benchmarking AI coding agents — built in the open at rossoctl.dev.
- moca - Dense, warm process-manager runtime for stateful AI coding agents — sessions resume exactly where they left off (TypeScript)
- cortex - The proxy layer between coding agents and everything they touch — see what your agent sends and pay less for it (Go) — install stats (estimated from GitHub download counts, not from user machines)
- eventing-lab - Signed-event agent wake path: a Kafka message wakes a coding agent and every step streams back as CloudEvents (Python)
- autobench - Automated benchmarking of agentic AI workloads — deploys the workload, collects per-task telemetry, publishes durable artifacts (Python)
- state-lab - Demo flows exploring state requirements for distributed agents, from durability to passability
- context-guru - Provider-agnostic context engineering for LLM agents — lower token cost and latency without sacrificing accuracy (Go)
- aiac - AI-based access control — translates natural-language policy into enforced permissions on OPA (Python)
- lab-data-governance - Data-aware governance: observes how data flows across agents, tools, and sessions, and derives interactions, classifications, and lineage (Python)
- context-service - Agent context infrastructure: durable storage and lifecycle management for agent state, memory, knowledge, workspaces, and artifacts (Go)
- automation - Automated org maintenance programs using a scanner/fixer pattern (Shell)
- repoman - RepoMan: automated repo-maintenance programs across enrolled GitHub repos (Shell)
- rossoctl - Installer, web console, and documentation (Python)
- .github - Project website and documentation (Docusaurus)
- 📖 Start with the CONTRIBUTING guide
- 👥 Meet the maintainers
- 🌐 Learn more at rossoctl.dev
- 💬 Or just come say hi in the Rossoctl Slack
