Agentically this could be done now, but it would still just be the tip of the iceberg
You're looking at quantum states of law and regulations, and that's just for one jurisdiction
So to make this a git repo, I guess you'd make the snapshots be the law plus markup for all the changes, and the diffs are the metachanges. For each snapshot S, you could generate a git repo whose snapshots are the law and the whose diffs are the changes in the law (as described by S). However, this probably isn't the best way to query S.
This makes sense for bills. It doesn't for laws. Once a bill is a law, it's the will of the Congress.
The legislative process is inherently collaborative in a way that's ordinarily undocumented. Who introduced an amendment and who supported it are entirely separate. Someone who introduced language can wind up not supporting it because of where the mix of trade-offs winds up.
At the level of a bill, knowing who wrote language matters. At the level of law, you should only care who voted for it. (I'm not even tackling the inherent subjectivity in tracing e.g. language in the CFR to specific language in a bill.)
even then you have three-line whips and numerous reasons to vote for / against things that are wrapped up in party politics and other stupid games
once something is passed; the intent of the writer, the intent of those that voted for it and the meaning of the words in the final legislation can all be wildly different.
If transparency is the goal, the entire codebase for this should be public domain and not profitable. Otherwise there's a profit motive to rewrite history.