Corpus preservation method
- Scope is restricted to public, unauthenticated material.
- Each page snapshot preserves raw HTML, extracted text, response headers, capture time, source URL, redirect target, and same-scope outgoing links.
- Crawls are bounded by explicit host/path allowlists. They are evidence snapshots, not mirrors of unrelated sites.
- Repository snapshots use shallow clones where possible and retain upstream URL plus captured commit.
- Gated Figma libraries or internal assets are inventoried by public reference; access is not inferred or bypassed.
manifest.md files describe provenance and local locations. They intentionally avoid capability scoring or recommendations during preservation.- Supporting sources remain in
/supporting and are not represented as peer systems.