Skip to main content

uv-installable wrapper for the vercel-labs/skills CLI with an embedded Node runtime

Project description

npx-skills

npx-skills is a uv-installable Python wrapper around the upstream skills CLI.

This allows you to use skills.sh from within Python projects without a global dependency on Node/npm and impose versioning constraints on npx skills on a project-by-project basis.

It does this by shipping a vendored Node runtime and the published JavaScript CLI payload, then forwards execution to the embedded runtime so users do not need a separate Node install.

Install

uv add npx-skills
skills -h

OR use with uvx

uvx npx-skills

Versioning

The npx-skills package version tracks the upstream skills npm module version one-to-one. Installing npx-skills==1.4.8 gives you exactly skills@1.4.8 — the CLI payload is baked into the wheel at build time, so pinning npx-skills in your pyproject.toml or lockfile pins the underlying JavaScript CLI too. A new upstream skills release on npm has no effect on an already-installed npx-skills; you only move when you explicitly bump npx-skills. Packaging-only republishes of an existing upstream version use PEP 440 post-releases (e.g. 1.4.8.post1).

How it works

Refresh Vendored Artifacts

The repository includes a vendoring script that copies:

  • the published skills package payload (package.json, bin/, dist/, README.md, ThirdPartyNoticeText.txt)
  • the package's runtime node_modules dependencies
  • the current platform's node binary

By default it will use a built package root if you pass --source, or fall back to the newest cached ~/.npm/_npx/*/node_modules/skills install.

uv run python scripts/vendor_upstream.py

Or point it at a specific built package root:

uv run python scripts/vendor_upstream.py --source /path/to/node_modules/skills

If you point --source at the upstream git repo, it must already contain dist/cli.mjs.

To rebuild only the embedded Node runtime from the payload already committed in this repo, point --source at the vendored package root:

uv run python scripts/vendor_upstream.py --source src/npx_skills/vendor/skills

Build

The wheel is intentionally platform-specific because it includes a native node binary.

uv build

Release Automation

GitHub Actions builds one wheel per native runner target:

  • Linux x64 via manylinux_2_28
  • Linux arm64 via manylinux_2_28
  • macOS x64
  • macOS arm64
  • Windows x64
  • Windows arm64

The workflow lives at .github/workflows/wheels.yml. It reuses the committed skills payload, re-vendors the Node runtime for the current runner architecture, runs the test suite, builds wheels with uv (cibuildwheel via uvx on Linux), smoke-tests the built wheel, and uploads wheel artifacts for manual release use.

This project intentionally publishes wheels only. An sdist would not be portable without re-vendoring the embedded Node runtime for the target platform.

Publishing runs through a separate manual workflow at .github/workflows/publish.yml. It invokes the wheel builder in the same workflow run, then publishes from GitHub after all target wheels succeed. The intended entrypoint is the npx-skills-dev helper:

uv run --package npx-skills-dev npx-skills-dev publish-to-pypi

Or via make:

make publish

Use --ref <branch> to dispatch the workflow from a specific branch and --dry-run to build and validate without publishing.

The GitHub publish job authenticates to PyPI using a project API token stored as the PYPI_API_TOKEN repository secret and consumed via UV_PUBLISH_TOKEN. The job runs in the pypi GitHub Actions environment, which can optionally be gated with required reviewers.

Upstream Sync Automation

A scheduled workflow at .github/workflows/auto-sync-upstream.yml runs daily (06:17 UTC) and keeps the vendored payload in lock-step with the upstream skills npm release. It is the mechanism that preserves the one-to-one version mapping described in the Versioning section above.

The job is implemented by the npx-skills-dev auto-sync-upstream command (packages/npx-skills-dev/src/npx_skills_dev/commands/auto_sync_upstream.py) and performs the following steps:

  1. Reads the currently vendored version from src/npx_skills/vendor/manifest.json (upstream.version).
  2. Fetches skills@latest from the npm registry.
  3. If the versions match (and --force was not passed) the job exits as a no-op.
  4. Otherwise it npm installs skills@<latest> into a temporary directory and runs scripts/vendor_upstream.py --source ... against it to refresh the vendored payload.
  5. Rewrites the version in pyproject.toml and src/npx_skills/__init__.py to match the new upstream version.
  6. Runs the unit test suite.
  7. Commits src/npx_skills/vendor, src/npx_skills/__init__.py, and pyproject.toml with the message Auto-sync upstream skills@<version> and pushes to master.
  8. Dispatches publish.yml against master, which builds wheels for every target platform and publishes them to PyPI.

The workflow can also be triggered manually via workflow_dispatch, with an optional force input that re-runs the full flow even when the upstream version is unchanged.

Layout

The vendored runtime is kept under src/npx_skills/vendor/:

vendor/
  manifest.json
  node/<platform>/node
  skills/
    package.json
    bin/cli.mjs
    dist/
    node_modules/

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distributions

No source distribution files available for this release.See tutorial on generating distribution archives.

Built Distributions

If you're not sure about the file name format, learn more about wheel file names.

npx_skills-1.5.1-py3-none-win_arm64.whl (30.1 MB view details)

Uploaded Python 3Windows ARM64

npx_skills-1.5.1-py3-none-win_amd64.whl (33.7 MB view details)

Uploaded Python 3Windows x86-64

npx_skills-1.5.1-py3-none-manylinux_2_28_x86_64.whl (45.5 MB view details)

Uploaded Python 3manylinux: glibc 2.28+ x86-64

npx_skills-1.5.1-py3-none-manylinux_2_28_aarch64.whl (44.7 MB view details)

Uploaded Python 3manylinux: glibc 2.28+ ARM64

npx_skills-1.5.1-py3-none-macosx_15_0_arm64.whl (37.3 MB view details)

Uploaded Python 3macOS 15.0+ ARM64

npx_skills-1.5.1-py3-none-macosx_10_13_universal2.whl (39.0 MB view details)

Uploaded Python 3macOS 10.13+ universal2 (ARM64, x86-64)

File details

Details for the file npx_skills-1.5.1-py3-none-win_arm64.whl.

File metadata

  • Download URL: npx_skills-1.5.1-py3-none-win_arm64.whl
  • Upload date:
  • Size: 30.1 MB
  • Tags: Python 3, Windows ARM64
  • Uploaded using Trusted Publishing? No
  • Uploaded via: uv/0.11.7 {"installer":{"name":"uv","version":"0.11.7","subcommand":["publish"]},"python":null,"implementation":{"name":null,"version":null},"distro":{"name":"Ubuntu","version":"24.04","id":"noble","libc":null},"system":{"name":null,"release":null},"cpu":null,"openssl_version":null,"setuptools_version":null,"rustc_version":null,"ci":true}

File hashes

Hashes for npx_skills-1.5.1-py3-none-win_arm64.whl
Algorithm Hash digest
SHA256 11a44329ae579cea85a5b373e23679fb3a2ebc5c96cfed38720bfa2c72316c6f
MD5 a1e3da43f9dcd0a44a248a2f8d34a82c
BLAKE2b-256 227ae4bf103e0211111404a8c2dc0ebc65e6d176dd287fdf11166d099efd980e

See more details on using hashes here.

File details

Details for the file npx_skills-1.5.1-py3-none-win_amd64.whl.

File metadata

  • Download URL: npx_skills-1.5.1-py3-none-win_amd64.whl
  • Upload date:
  • Size: 33.7 MB
  • Tags: Python 3, Windows x86-64
  • Uploaded using Trusted Publishing? No
  • Uploaded via: uv/0.11.7 {"installer":{"name":"uv","version":"0.11.7","subcommand":["publish"]},"python":null,"implementation":{"name":null,"version":null},"distro":{"name":"Ubuntu","version":"24.04","id":"noble","libc":null},"system":{"name":null,"release":null},"cpu":null,"openssl_version":null,"setuptools_version":null,"rustc_version":null,"ci":true}

File hashes

Hashes for npx_skills-1.5.1-py3-none-win_amd64.whl
Algorithm Hash digest
SHA256 7a6b81c004d8cb5ac5fd0910b012c406c8d0b8e1dfb72eeee19c838368b863b4
MD5 c7a7bce748a84e15ace796b50df6729c
BLAKE2b-256 bef557de01911cf8fb91738e1a10e3530ecdcef0b8c0595c7b3a65c04b86c637

See more details on using hashes here.

File details

Details for the file npx_skills-1.5.1-py3-none-manylinux_2_28_x86_64.whl.

File metadata

  • Download URL: npx_skills-1.5.1-py3-none-manylinux_2_28_x86_64.whl
  • Upload date:
  • Size: 45.5 MB
  • Tags: Python 3, manylinux: glibc 2.28+ x86-64
  • Uploaded using Trusted Publishing? No
  • Uploaded via: uv/0.11.7 {"installer":{"name":"uv","version":"0.11.7","subcommand":["publish"]},"python":null,"implementation":{"name":null,"version":null},"distro":{"name":"Ubuntu","version":"24.04","id":"noble","libc":null},"system":{"name":null,"release":null},"cpu":null,"openssl_version":null,"setuptools_version":null,"rustc_version":null,"ci":true}

File hashes

Hashes for npx_skills-1.5.1-py3-none-manylinux_2_28_x86_64.whl
Algorithm Hash digest
SHA256 ccb1742c5ab098fc7692bc53a72eb7d39f863cba6a7b3acfe82dda2d5995553e
MD5 581ad80b0f3c949c1f09ac4dc831bb6a
BLAKE2b-256 eebb5d9a7713a50365c21c83583b2477a4239925e5756125c24571f3c2f94e20

See more details on using hashes here.

File details

Details for the file npx_skills-1.5.1-py3-none-manylinux_2_28_aarch64.whl.

File metadata

  • Download URL: npx_skills-1.5.1-py3-none-manylinux_2_28_aarch64.whl
  • Upload date:
  • Size: 44.7 MB
  • Tags: Python 3, manylinux: glibc 2.28+ ARM64
  • Uploaded using Trusted Publishing? No
  • Uploaded via: uv/0.11.7 {"installer":{"name":"uv","version":"0.11.7","subcommand":["publish"]},"python":null,"implementation":{"name":null,"version":null},"distro":{"name":"Ubuntu","version":"24.04","id":"noble","libc":null},"system":{"name":null,"release":null},"cpu":null,"openssl_version":null,"setuptools_version":null,"rustc_version":null,"ci":true}

File hashes

Hashes for npx_skills-1.5.1-py3-none-manylinux_2_28_aarch64.whl
Algorithm Hash digest
SHA256 b0fcac3a62192d1d9bb902b29300efe436bbfdca59355d24ef56e4d7cfdcc5fb
MD5 203bd0f8d79e0107e0dc6f6ecdd41a88
BLAKE2b-256 16bdfb69701f76175fe74386ba9eb9d98d4d83c5ad13bf4d62f33412b160fe7a

See more details on using hashes here.

File details

Details for the file npx_skills-1.5.1-py3-none-macosx_15_0_arm64.whl.

File metadata

  • Download URL: npx_skills-1.5.1-py3-none-macosx_15_0_arm64.whl
  • Upload date:
  • Size: 37.3 MB
  • Tags: Python 3, macOS 15.0+ ARM64
  • Uploaded using Trusted Publishing? No
  • Uploaded via: uv/0.11.7 {"installer":{"name":"uv","version":"0.11.7","subcommand":["publish"]},"python":null,"implementation":{"name":null,"version":null},"distro":{"name":"Ubuntu","version":"24.04","id":"noble","libc":null},"system":{"name":null,"release":null},"cpu":null,"openssl_version":null,"setuptools_version":null,"rustc_version":null,"ci":true}

File hashes

Hashes for npx_skills-1.5.1-py3-none-macosx_15_0_arm64.whl
Algorithm Hash digest
SHA256 a69fc37e23b241375b69c9e7abdd808e86a743fd8c8728597a0d9b288a6cc56b
MD5 820e4cd313979276dda42bc2e1c84e22
BLAKE2b-256 4ceb1d0aa51fb2a5abc738724dde8714ec141a7467c50bf8db64b96d1f009e59

See more details on using hashes here.

File details

Details for the file npx_skills-1.5.1-py3-none-macosx_10_13_universal2.whl.

File metadata

  • Download URL: npx_skills-1.5.1-py3-none-macosx_10_13_universal2.whl
  • Upload date:
  • Size: 39.0 MB
  • Tags: Python 3, macOS 10.13+ universal2 (ARM64, x86-64)
  • Uploaded using Trusted Publishing? No
  • Uploaded via: uv/0.11.7 {"installer":{"name":"uv","version":"0.11.7","subcommand":["publish"]},"python":null,"implementation":{"name":null,"version":null},"distro":{"name":"Ubuntu","version":"24.04","id":"noble","libc":null},"system":{"name":null,"release":null},"cpu":null,"openssl_version":null,"setuptools_version":null,"rustc_version":null,"ci":true}

File hashes

Hashes for npx_skills-1.5.1-py3-none-macosx_10_13_universal2.whl
Algorithm Hash digest
SHA256 168592a095993b58fc386e7617a21cef4c206908766939e815c701036b3c3a6b
MD5 4d863672debbfa7518f502551ac79a90
BLAKE2b-256 8476b3fbe1bbbeea063d4f825ce4763a0cf5fb014110df27b46321405a6c6683

See more details on using hashes here.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Depot Continuous Integration Fastly CDN Google Download Analytics Pingdom Monitoring Sentry Error logging StatusPage Status page