Using pre-commit as a polyglot task runner: elegant or kludgy?

Package maintainers: does your Makefile or justfile delegate tool calls to pre-commit (e.g., for your lint all targets)?

I see a lot of modern repos doing it this way now. On one hand, I can see the elegance of it—pre-commit has basically evolved into an execution engine that manages isolated polyglot tool environments.

But it still just feels weird to me, almost like a layer violation. Maybe it's just because my mental model still views pre-commit strictly as a git-hook manager rather than what it has actually become.

One huge benefit, I guess, of having linting and quality tools delegated this way is parity: your CI pipeline, your pre-commit hooks, and your manual justfile targets all run the exact same tools in the exact same way.

It also solves the polyglot problem. Modern dev dependencies can't all be managed by one tool (like uv, pip, or pnpm) because some are Node, some are Python, some are Go, etc.

Curious to hear how others are approaching this. Any strong reasons for / against delegating to pre-commit for your task runners vs. keeping them strictly decoupled?