liberate
Installation
SKILL.md
Liberate a website
What you produce: a directory that runs on its own — every retained route, its CSS, media, fonts, navigation, same-page anchors, and responsive behavior, with references rewritten to point inside the copy.
HTML is the contract. That directory is the deliverable, not an intermediate step toward a platform.
The CLI does the work. Your job is to run it, read what it reports, verify the result, and tell the operator what they got.
The whole surface
| Command | What it does |
|---|---|
data-liberation inspect <url> |
Bounded read-only source assessment with explicit coverage and unknowns |
data-liberation <url> |
Detect the platform, discover routes, liberate every one, write the site |
data-liberation compare <run-dir> |
Verify the copy against its live source. This is the acceptance gate |
data-liberation publish <run-dir> --to <target> |
Put the copy on a live URL |