Skip to content

Draft — not approved. A working draft by UASK. It may change without notice.

Status: review

UASK Engineering Hub

The UASK Engineering Hub is an open, versioned engineering knowledge base for developers and AI coding agents who build software products of different kinds.

Every rule has one source, a version, a status and evidence, so a project can pin exactly the version it follows. Unlike scattered wikis, books and ad-hoc prompt files, the same knowledge is readable by people and executable by agents.

Define once — Learn, Build, Verify, Reuse everywhere.

The standards that live here: HIF — Human Interface Framework, and SCF — System Construction Framework.

Sections

Section What it holds
Engineering Core platform-independent standards and practices
Platform profiles additions for one platform, extending core items
Playbooks end-to-end scenarios that combine items, templates and skills
Learn tutorials and explanations, without rules
Glossary one term, one definition

All items are in review: they are open for reading and comment, and may still change before the first release.

How to add an item

  1. Pick the section and kind: a rule set or practice goes to content/core/<domain>/, a platform addition to content/profiles/<profile>/, a tutorial to content/learn/, a term to content/glossary/.
  2. Copy templates/knowledge-item.md to <slug>.md and fill in the front matter. The id is <domain prefix>.<slug> (for example arch.adr); prefixes are listed in schemas/knowledge-item.schema.json. A new item starts at version 0.1.0 or 1.0.0 with status draft.
  3. Write the body in the fixed order: Summary, Rules, Guidance, Explanation, Examples, Sources. Upper-case MUST, SHOULD and MAY appear only under Rules, one rule per - **Rn.** line.
  4. Cite at least one primary source with its licence (SPDX id) and access date. Write in your own words. Material adapted from CC BY-SA sources goes only to content/adapted/; NC and ND sources are linked, never copied.
  5. Run python3 tools/validate_content.py --standards and bash tests/run.sh, then open a pull request with status review.