Good First Issues in V

No security policy exists in the repo despite wrapping native C libraries (OpenBLAS, LAPACKE, CUDA, Vulkan, HDF5, MPI). Task Create SECURITY.md with: Supported versions (latest release) Reporting process (email or GitHub Security Advisory) Response timeline expectation File New SECURITY.md

compute/ is the recommended dispatch surface for backend-agnostic GPU computing but has no README. All other GPU-related modules (vcl, cuda, vulkan) have READMEs. Task Create compute/README.md explaining: The compute abstraction layer How it dispatches to vcl/cuda/vulkan backends Available operations (element-wise, BLAS, reduce) Basic usage example Verification README links correctly to backend-specific docs

errors/errno.v, errors/errors.v, errors/strerrno.v have no test coverage. Task Add errors/errno_test.v mapping known errno integers to their string representations, and verifying error code constants are consistent. Verification v test vsl/errors passes

Three articles in docs/advanced/ are stubs: 01-mpi-parallel.md — 308 bytes 03-hdf5-io.md — 285 bytes 04-library-integration.md — 502 bytes Examples for each topic already exist (mpi, inout, external library use). Task Write real walkthroughs for each article using the existing examples as reference. Verification Follow the steps in each article to run the example Articles follow the depth of 02-opencl-gpu.md

Three modules have minimal READMEs: dist/README.md — 112 bytes noise/README.md — 2 sentences roots/README.md — 1 sentence Task Expand each README with: Module purpose and capabilities Public API listing with usage snippets Links to examples and related docs Use metrics/README.md as reference for the expected quality level. Verification Each README covers the module's full public API