Run the Kaappi unit suite under -Dgc-stress=true on a real x86-64 Linux machine using a DigitalOcean droplet with a 3-hour lifetime. Use when the user asks to stress-test on Linux, run the gc-stress suite on real hardware, verify GC rooting under stress on x86-64, or stress test on DigitalOcean. Complements /do-linux-test (normal suite, ~1-hour droplet) — this variant runs the collection-per-allocation build, which takes hours, not minutes.
Post a release announcement to the kaappi org Announcements forum (github.com/orgs/kaappi/discussions). Use when the user asks to announce a release, post a release announcement, tell the community about a version, or share a release on the forum. Takes a release tag such as v0.22.0.
Run the KEP-0002 Phase 7 gate-campaign benchmark (benchmarks/gate/) on a real x86-64 Linux reference machine using a DigitalOcean droplet, producing the statistically rigorous Kalibera-Jones dataset a KEP acceptance gate reads. Use when a KEP's acceptance gate needs its Linux/second reference-machine dataset, when a documented revisit trigger fires and the gate needs re-running, or when a new performance-gated KEP wants the same dual-machine benchmark protocol. Distinct from /do-stress-test (gc-stress unit suite) and /do-linux-test (normal build + Scheme test suite) — this runs a multi-hour statistical benchmark harness, not a test suite, and needs ≥8 physical cores.
Run Kaappi build and full test suite on a real x86-64 Linux machine using a DigitalOcean droplet. Use when the user asks to test on x86-64 Linux, run Linux tests on real hardware, verify x86-64 compatibility with full Scheme test suites, or test on DigitalOcean. Complements /linux-test (podman-based) by providing real x86-64 hardware instead of emulation.
Group open GitHub issues into parallel sets for concurrent Claude Code sessions. Use when the user asks to batch issues, plan parallel work, triage open issues for parallel fixing, or create issue sets. Accepts an optional label (or multiple comma-separated labels) to restrict which issues are considered, and an optional second argument of comma-separated issue numbers to skip.
Power on a local UTM virtual machine and run Kaappi builds/tests on it over SSH. Use when the user asks to turn on a VM and test/build, to test or build on a UTM VM, on FreeBSD/OpenBSD/NetBSD/Windows, on the s390x or ppc64le Alpine VMs, on a real BSD or Windows box, or to verify the current branch on one of the local VMs. Complements /linux-test (podman) and /do-linux-test (DigitalOcean) with the local UTM fleet.
Run Kaappi build and tests on Linux using podman. Tests aarch64 (native build + unit tests + R7RS), x86_64 (cross-compile + JIT tests), and riscv64 (cross-compile + interpreter tests). Use when the user asks to test on Linux, verify cross-platform build, or check Linux compatibility.