packstruct docs

Fleet and enterprise

The moment more than one team writes packs, “it loads on my machine” stops being good enough. This page is the adoption path for a fleet: how the spec scales, how to gate CI on it, and how to bring a problem to the issue tracker. The branded version is at packstruct.creadev.org/enterprise.

What enterprise actually needs

An adoption path that scales

  1. Audit what you have. Run npx packstruct lint --deep over each repo’s packs/ directory. You get a nameable list of every manifest that points at nothing, and every id collision.
  2. Pin the spec you adopt. Add packstruct as an exact devDependency and a receipt row: version, the commit it was cut against, the lint result, the date.
  3. Gate CI on the lanes. lint and audit-conformance both exit 0 clean, 1 violations, 2 usage. Wire them as a required check.
  4. Declare your conformance. Carry implements: packstruct@1 in your docs and serve your verdicts with the family-status@1 dialect.
# 1. audit
npx packstruct lint packs/ --deep

# 3. the two CI gates (both exit 0 clean / 1 violations / 2 usage)
npx packstruct lint packs/ --deep
npx packstruct audit-conformance . --loader "node loader.mjs {packs}"

Every step above is already how the live hosts in the Showcase run. You are adopting a lane with receipts, not a proposal.

Open source, and staying that way

packstruct is MIT-licensed. The spec, the CLI, and the harness are all open, and development happens in the open through pull requests with their gates run. The CLI ships with zero runtime dependencies, so adopting it adds nothing to your supply chain. Published on npm as packstruct.

Bring us a problem

The fastest way to shape where the spec goes is to file what you need. These are the prefilled issue forms - each opens GitHub’s own composer with the right template, which you review and submit there:

Prefer to see what is already open? Browse the issue tracker first; a good issue is a searchable receipt that helps the next adopter.