메뉴
Run RCLI smoke and e2e against a pinned C++ desktop kit on macOS and Windows. Use when verifying a kit pin, a Windows link fix, MLX host, or when the user asks to e2e-test rcli.
기본적으로 소스를 먼저 확인하는 Prompt가 선택됩니다. 직접 명령으로 전환하거나 로컬 사본을 다운로드할 수도 있습니다.
설치 여부를 결정하기 전에 SKILL.md와 SkillsMP에 표시된 보조 파일을 읽어 보세요.
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
직접 명령은 검토 Prompt를 거치지 않습니다. 실행하기 전에 소스를 확인하세요.
npx skills add https://github.com/RunanywhereAI/RCLI --skill rcli-e2e명령은 한 줄로 유지됩니다. 복사하기 전에 가로로 스크롤해 전체 내용을 확인하세요.
로컬 사본을 원하시나요? SkillsMP에서 현재 제공할 수 있는 파일을 다운로드하세요.
RCLI layering and command-authoring rules. Use when adding or changing CLI commands, deciding where logic belongs, touching proto/ABI boundaries, or when the user mentions bootstrap, rac_*, thin commands, or kit vs CLI.
Bump RCLI's pinned C++ desktop kit and IDL schema lock. Use when consuming a new RunAnywhere kit, sdk-pin.cmake is stale, SCHEMA_LOCK mismatches, or the user mentions find_package(RunAnywhere), fetch-kit, or pin bump.
Cut an RCLI product release independent of the SDK kit version. Downloads the pinned C++ desktop kit, builds bottles, signs, stamps Formula/rcli.rb, and publishes a GitHub Release. Use when tagging rcli, shipping bottles, or updating the Homebrew formula.
SKILL.md 표시 중
| name | rcli-e2e |
| description | Run RCLI smoke and e2e against a pinned C++ desktop kit on macOS and Windows. Use when verifying a kit pin, a Windows link fix, MLX host, or when the user asks to e2e-test rcli. |
Always test the product binary against a released or local kit, never by compiling SDK source into this tree.
bash scripts/fetch-kit.sh macos-arm64 /tmp/kit
cmake -B build -G Ninja -DCMAKE_BUILD_TYPE=Release -DCMAKE_PREFIX_PATH=/tmp/kit
cmake --build build -j "$(sysctl -n hw.logicalcpu)"
ctest --test-dir build --output-on-failure
bash scripts/e2e.sh ./build/rcli-cxx
Optional real generate: RCLI_E2E_MODEL=qwen3-0.6b bash scripts/e2e.sh ./build/rcli-cxx.
Apple shipping binary (MLX host):
bash scripts/build-mlx.sh
bash scripts/smoke-mlx.sh
C++ rcli-cxx cannot exercise MLX (needs Swift host callbacks). That is
expected, not a skip to paper over.
Host cmake/ninja on PATH may be Python ARM64 — use VS x64 after vcvarsall x64.
Kit is /MD. Do not force /MT.
vcvarsall.bat x64
bash scripts/fetch-kit.sh windows-x64 C:\kit
cmake -B build -G Ninja -DCMAKE_BUILD_TYPE=Release -DCMAKE_PREFIX_PATH=C:/kit
cmake --build build --config Release
ctest --test-dir build --output-on-failure -C Release
bash scripts/e2e.sh ./build/rcli.exe
Copy onnxruntime.dll from the kit third_party/ next to rcli.exe if smoke
cannot load ORT. Do not tarball swift/ (multi-GB).
CI already runs unit + modelless e2e on macos-14 and windows-2022. A pin bump
is not done until those are green. Linux e2e is optional (scripts/e2e-linux.sh).