crsh/prereg/lint-prereg-skeleton

Lint Prereg Skeleton

Applies consistent Markdown formatting to skeleton.Rmd preregistration templates. The canonical references are vantveer_prereg and secondary_data_analysis_prereg.

See detailed rules with examples.

When to Use

• Adding a new template to inst/rmarkdown/templates/
• Editing an existing skeleton and wanting to restore consistent style
• Auditing all templates for formatting consistency

Procedure

1. Identify the target file(s)

If an argument is given, format that file. Otherwise format every file matching:

inst/rmarkdown/templates/*/skeleton/skeleton.Rmd

2. Apply the lint script

Run .github/skills/lint-prereg-skeleton/lint_skeleton.R on the target file(s):

Rscript .github/skills/lint-prereg-skeleton/lint_skeleton.R <path/to/skeleton.Rmd>

The script automates Rules 1–11 and 13–15 (see formatting-rules.md): trailing whitespace, comment delimiter spacing, blank-line removal between headings and comments, prompt-text normalisation, section spacing, \newpage padding, bullet style, sentence-case headings, and comment block internal formatting (paragraph joining, blank-line separation, leading whitespace removal).

3. Verify script output

Review the diff produced by the script. Check that every automated rule has been applied correctly and consistently — the script may mis-classify edge cases (e.g. bold sub-labels inside ## sections, slash-compound headings). If a rule was applied incorrectly:

• Evaluate whether the script can be improved to handle the case.
• If so, update lint_skeleton.R and re-run before continuing.
• Otherwise, revert the incorrect change manually.

4. Apply remaining rules manually

Work through the rules the script does not cover, and fix any edge cases the script missed:

• Rule 12 — Spelling and grammar: fix clear spelling errors and obvious grammar errors in headings and comment text. Preserve original wording and intent; do not paraphrase. Leave examples, R code, LaTeX, and YAML untouched.
• Rule 16 — Unlabeled bullet lists: Rules 13–15 join unlabeled items into paragraphs. After the script runs, inspect joined paragraphs to identify hidden lists and add - markers manually.
• Missing prompts: add Enter your response here. to any Type A section that lacks one.
• Encoding issues: correct any garbled characters (e.g. Í → ').

5. Verify the YAML front matter and References section

Do not alter:

• The YAML front matter block (--- ... ---)
• The # References section and the LaTeX spacing commands that follow it
• R inline code (` r ... `)
• The prose content of comment blocks — only list markers, indentation, whitespace, question-number prefixes, and clear spelling/grammar errors within them may change
• Blank lines within a comment block that separate paragraphs — those are intentional
• Markdown tables and fenced code blocks

6. Final check

After writing, verify:

• No line has trailing whitespace.
• Every <!-- adjacent to text has a space after it; every --> adjacent to text has a space before it.
• Every heading is immediately followed by a comment or content with no blank line between them.
• Type A sections: --> → 1 blank → prompt → 2 blank → next heading.
• Type B sections: choices → 2 blank → next heading (no prompt).
• Type C sections: --> → 2 blank → next heading (no prompt).
• Every \newpage has 2 blank lines before and after.
• All lists use N. or - markers with correct indentation, everywhere.
• All heading text (except # References) is in sentence case.