quilt-core is Quilt's open-source runtime: container ops, icc networking, oci image workflow, sync/state engine, and local operational tooling.
Included in OSS:
- gRPC daemon (
quilt) and CLI (cli) - Linux container runtime primitives (namespaces, cgroups, process supervision)
- SQLite-backed sync/state engine
- ICC networking stack (bridge, veth, DNS manager, firewall/netlink path)
- Image/registry workflows and runtime support modules
- Volume and metrics paths
- Fozzy scenarios and verification tooling
- NOTE: For native Kubernetes-like cluster management, use quiltc
- Linux host (cgroups + namespaces enabled)
- Rust toolchain (edition 2021; recent stable recommended)
protoc(for gRPC/protobuf codegen)- common build tools (
gcc,pkg-config)
cargo build --releaseBinaries:
target/release/quilttarget/release/clitarget/release/minit
Start daemon:
./target/release/quiltBasic CLI flow:
./target/release/cli --help
./target/release/cli list
./target/release/cli imagesOptional local stack:
docker compose up --buildFozzy is the primary verification path in this repo.
Recommended full gate:
fozzy full \
--scenario-root tests \
--seed 1337 \
--doctor-runs 5 \
--fuzz-time 2s \
--explore-steps 200 \
--explore-nodes 3 \
--allow-expected-failures \
--require-topology-coverage . \
--topology-min-risk 60 \
--topology-profile pedantic \
--jsonAdditional script tests are in tests/ (container, ICC, sync, volume, stress, diagnostics).
src/
main.rs # gRPC daemon
cli/ # CLI client
daemon/ # runtime + system integration
sync/ # SQLite state/scheduling/orchestration
icc/ # networking + messaging
image/ # OCI image handling
registry/ # registry client/auth flows
usage/ # usage event tracking
utils/ # shared runtime utilities
proto/ # gRPC proto definitions
scripts/ # install/migration/dev helpers
tests/ # shell + fozzy test scenarios
- Keep changes scoped and production-safe for self-hosting.
- Validate with Fozzy full gate before PR.
- Avoid introducing cloud-only HTTP/auth/subscription/serverless surface into
quilt-core.
MIT OR Apache-2.0