I think we could benefit from extracting some data from the code itself and moving it to a repository that's cloned on first run, like the community index. This would also allow updating this information comfortably without needing to push binary updates. Candidate data is:

• Templates for new crates
• List of trusted hosting sites
• Help texts?

Although we should carefully consider how this would affect pure off-line use, which is currently doable with private indexes or just to kickstart crates. Maybe a fallback version could be packaged as compiled-in resources with awsres.