Problem
The mashlib npm package currently has 590 published versions totaling an estimated ~12GB on the npm registry (each build is ~21MB unpacked).
The growth has accelerated sharply:
| Year |
Versions published |
| 2023 |
23 |
| 2024 |
10 |
| 2025 |
90 |
| 2026 (Jan–Feb only) |
109 |
That's 199 versions in ~14 months, compared to 10 in the entirety of 2024.
Risk
npm monitors package storage and has taken action against packages for excessive registry usage. With 14 maintainers listed on this package, a policy action would affect everyone.
Suggested fixes
- Add a version-exists check to CI so it doesn't publish if the version already exists on npm
- Stop publishing prerelease/hash versions (e.g.
2.1.4-test.0, 3.0.1-0dd33a6) to the public registry — use a local artifact or GitHub Packages instead
- Consider deprecating the hundreds of unused prerelease versions to signal they are not intended for consumption
- Review the CI pipeline to ensure publishes only happen on tagged releases from main
Context
This is a shared infrastructure concern that affects all 14 npm maintainers. The goal is to prevent further storage growth and reduce the risk of npm taking action against the package.
cc @jeswr @timea-solid
Problem
The
mashlibnpm package currently has 590 published versions totaling an estimated ~12GB on the npm registry (each build is ~21MB unpacked).The growth has accelerated sharply:
That's 199 versions in ~14 months, compared to 10 in the entirety of 2024.
Risk
npm monitors package storage and has taken action against packages for excessive registry usage. With 14 maintainers listed on this package, a policy action would affect everyone.
Suggested fixes
2.1.4-test.0,3.0.1-0dd33a6) to the public registry — use a local artifact or GitHub Packages insteadContext
This is a shared infrastructure concern that affects all 14 npm maintainers. The goal is to prevent further storage growth and reduce the risk of npm taking action against the package.
cc @jeswr @timea-solid