aspens — Scoped Skills from Your Import Graph

Main

• Replace the monolith: generate multiple small skill files so the agent reads the right context only when it touches that domain.

• Use doc impact as a freshness/coverage gate before trusting your context in CI or during high-risk refactors.

• Keep drift low by running doc sync after changes (or install the post-commit hook) so skills match current imports and domains.

Source-backed notes

• README describes skills as ~35-line markdown files that activate when the agent edits matching domains.
• README documents aspens doc init --recommended and aspens doc impact as the first verification loop.
• README includes aspens doc sync and an --install-hook option for post-commit auto-sync.

FAQ

• Do I need to install it globally?: No — README shows npx aspens doc init --recommended as a no-install path.
• What triggers a skill?: README explains skills activate when the agent touches matching domains/files, instead of loading one huge CLAUDE.md.
• How do I keep it up to date?: Run aspens doc sync after changes or install the post-commit hook via --install-hook.