fix: remove need for GH token when linting for conda-forge

[beckermr]

Checklist

• Added a news entry
• Regenerated schema JSON if schema altered (python conda_smithy/schema.py)

This PR

• removes the need for a GitHub token in order to run the linter for conda-forge locally
• moves staged-recipes specific lints to staged-recipes (xref: feat: unify linting staged-recipes#27590)

The idea here is three-fold:

• separate staged-recipes code (which is more adhoc and moves quickly) from code for feedstocks (which moves at a slower pace) (see, e.g. the fix we had to push in PR fix: remove lint for changes to staged-recipes example #2038 due to a change in staged-recipes - we removed the lint there in favor of staged-recipes code and this PR follows that same logic to its conclusion.)
• enable staged-recipes maintainers to help maintain the linter in ways that help them review recipes more efficiently without having to make a smithy PR and release (see, e.g., the custom github actions added currently for additional lints not covered by smithy).
• remove the need for a GH token when linting for conda-forge in order to enable linting in isolated environments (as in conda-forge-feedstock-ops) with no permissions / tokens to enhance our security

The staged-recipes lints have been added to staged-recipes in PR conda-forge/staged-recipes#27590.

xref: #2043

[beckermr]

pre-commit.ci autofix

[beckermr]

@conda-forge/core this PR is ready for review!

[beckermr]

bump here @conda-forge/core