AI Policy Authoring Kit

ArchLinterNet is YAML-first, which makes it suitable for AI-assisted architecture governance. Agents can inspect a repository, propose a policy, and review policy changes without writing custom architecture tests.

This public AI section is for product usage: authoring and reviewing ArchLinterNet policies. It is safe to publish as part of the GitHub Pages documentation site.

Internal repository-maintenance instructions, backlog governance, issue-writing rules, and implementation planning notes are not part of the published product documentation. They remain in GitHub Markdown files under internal repository locations.

Start here

• Agent Guide — investigation workflow before editing a policy.
• Policy Authoring Guide — YAML authoring rules.
• Capabilities — supported contract families and limits.
• Policy Review Checklist — review checks before opening a PR.

Machine-readable references

• schema/dependencies.arch.schema.json describes supported YAML structure.
• archlinternet.capabilities.json describes supported contract families, matching semantics, and limits.

Guardrail

Start from real assemblies and namespaces. Do not invent ideal architecture words, unsupported YAML fields, fake contract families, or broad ignored_violations entries to make a policy pass.

If a desired capability is not documented in the public product docs or schema, track it as a backlog proposal instead of authoring unsupported YAML.