.gitattributes

# Auto detect text files and ensure proper line endings
* text=auto

# Force LF (Unix) line endings for all common text-based files
*.php text eol=lf
*.js text eol=lf
*.css text eol=lf
*.scss text eol=lf
*.json text eol=lf
*.md text eol=lf
*.html text eol=lf
*.txt text eol=lf
*.xml text eol=lf
*.yml text eol=lf
*.yaml text eol=lf

# The changelog is append-only and almost every pull request adds an entry to
# the same short `[Unreleased]` list, so two open pull requests edit the same
# lines. A plain merge then conflicts, and the losing pull request has to have
# `develop` merged in by hand before it can pass the required status checks.
#
# `union` resolves that automatically by keeping both sides' additions, which
# is the correct merge for a file nobody edits in place. A genuine duplicate is
# still caught: the changelog validator's CHK_UNIQUE_CONTENT rule flags entries
# more than 90% similar, and
# `scripts/validation/__tests__/changelog-merge-strategy.test.js` fails if this
# line is removed or changed.
CHANGELOG.md merge=union

# Treat images and fonts as binary
*.png binary
*.jpg binary
*.jpeg binary
*.gif binary
*.svg binary
*.woff binary
*.woff2 binary
*.ttf binary

# --- Linear review categories ---------------------------------------------
#
# Linear reads these attributes to group the files in a pull request's diff
# ("File type" grouping) and to count implementation lines separately from
# tests and documentation. Linear counts implementation as "everything not
# claimed as something else", so a file that carries two categories is
# counted as implementation and the split stops working.
#
# Supported values:
#
#   review-implementation   the change itself
#   review-test             test code
#   review-documentation    prose and docs
#   review-generated        machine-generated or vendored output
#   review-agent-guidance   instructions for humans and AI agents
#   review-localization     translations (unused today; no locale trees)
#   review-assets           images and fonts
#
# WHY EVERY RULE IS THIS VERBOSE
#
# Git resolves each attribute name independently and the last matching rule
# wins per name. So negating a category only holds until some later rule sets
# it again -- negating inside a rule is not enough on its own. The invariant
# that actually holds is: the last matching rule sets one category and clears
# all six others, so it fully determines the file's state and every file lands
# in exactly one category. The rules below therefore look repetitive by design.
#
# Because each rule is self-contained, order only decides which category wins
# where patterns overlap, so the blocks run broad to specific and the most
# specific (tests) go last.
#
# Note `* text=auto` stays at the top of this file: it is the existing
# baseline that makes GitHub detect text and normalise line endings, and every
# rule here assumes it has already been applied.

# Baseline: everything is implementation until a rule below claims it.
* review-implementation

# Generated output and installed dependencies: real diff noise, no review value.
coverage/** review-generated -review-implementation -review-test -review-documentation -review-agent-guidance -review-localization -review-assets
node_modules/** review-generated -review-implementation -review-test -review-documentation -review-agent-guidance -review-localization -review-assets
website/dist/** review-generated -review-implementation -review-test -review-documentation -review-agent-guidance -review-localization -review-assets
website/.astro/** review-generated -review-implementation -review-test -review-documentation -review-agent-guidance -review-localization -review-assets
logs/** review-generated -review-implementation -review-test -review-documentation -review-agent-guidance -review-localization -review-assets
.github/reports/** review-generated -review-implementation -review-test -review-documentation -review-agent-guidance -review-localization -review-assets
metrics-artifacts/** review-generated -review-implementation -review-test -review-documentation -review-agent-guidance -review-localization -review-assets
package-lock.json review-generated -review-implementation -review-test -review-documentation -review-agent-guidance -review-localization -review-assets
*.min.js review-generated -review-implementation -review-test -review-documentation -review-agent-guidance -review-localization -review-assets
*.min.css review-generated -review-implementation -review-test -review-documentation -review-agent-guidance -review-localization -review-assets

# Images and fonts, plus the folders that hold them.
assets/** review-assets -review-implementation -review-test -review-documentation -review-generated -review-agent-guidance -review-localization
website/public/assets/** review-assets -review-implementation -review-test -review-documentation -review-generated -review-agent-guidance -review-localization
website/public/fonts/** review-assets -review-implementation -review-test -review-documentation -review-generated -review-agent-guidance -review-localization
website/src/assets/** review-assets -review-implementation -review-test -review-documentation -review-generated -review-agent-guidance -review-localization
*.png review-assets -review-implementation -review-test -review-documentation -review-generated -review-agent-guidance -review-localization
*.jpg review-assets -review-implementation -review-test -review-documentation -review-generated -review-agent-guidance -review-localization
*.jpeg review-assets -review-implementation -review-test -review-documentation -review-generated -review-agent-guidance -review-localization
*.gif review-assets -review-implementation -review-test -review-documentation -review-generated -review-agent-guidance -review-localization
*.svg review-assets -review-implementation -review-test -review-documentation -review-generated -review-agent-guidance -review-localization
*.webp review-assets -review-implementation -review-test -review-documentation -review-generated -review-agent-guidance -review-localization
*.ico review-assets -review-implementation -review-test -review-documentation -review-generated -review-agent-guidance -review-localization
*.woff review-assets -review-implementation -review-test -review-documentation -review-generated -review-agent-guidance -review-localization
*.woff2 review-assets -review-implementation -review-test -review-documentation -review-generated -review-agent-guidance -review-localization
*.ttf review-assets -review-implementation -review-test -review-documentation -review-generated -review-agent-guidance -review-localization
*.eot review-assets -review-implementation -review-test -review-documentation -review-generated -review-agent-guidance -review-localization

# Prose. Broad on purpose: it runs before the test, agent-guidance and
docs/** review-documentation -review-implementation -review-test -review-generated -review-agent-guidance -review-localization -review-assets
cookbook/** review-documentation -review-implementation -review-test -review-generated -review-agent-guidance -review-localization -review-assets
website/README.md review-documentation -review-implementation -review-test -review-generated -review-agent-guidance -review-localization -review-assets
*.md review-documentation -review-implementation -review-test -review-generated -review-agent-guidance -review-localization -review-assets

# Translations. No locale trees exist yet; the bucket stays reserved so the
**/languages/** review-localization -review-implementation -review-test -review-documentation -review-generated -review-agent-guidance -review-assets
**/locales/** review-localization -review-implementation -review-test -review-documentation -review-generated -review-agent-guidance -review-assets
**/i18n/** review-localization -review-implementation -review-test -review-documentation -review-generated -review-agent-guidance -review-assets
**/*.po review-localization -review-implementation -review-test -review-documentation -review-generated -review-agent-guidance -review-assets
**/*.pot review-localization -review-implementation -review-test -review-documentation -review-generated -review-agent-guidance -review-assets
**/*.mo review-localization -review-implementation -review-test -review-documentation -review-generated -review-agent-guidance -review-assets

# Agent and human guidance. After the *.md rule so these win over prose.
AGENTS.md review-agent-guidance -review-implementation -review-test -review-documentation -review-generated -review-localization -review-assets
CLAUDE.md review-agent-guidance -review-implementation -review-test -review-documentation -review-generated -review-localization -review-assets
GEMINI.md review-agent-guidance -review-implementation -review-test -review-documentation -review-generated -review-localization -review-assets
.github/custom-instructions.md review-agent-guidance -review-implementation -review-test -review-documentation -review-generated -review-localization -review-assets
.github/copilot-instructions.md review-agent-guidance -review-implementation -review-test -review-documentation -review-generated -review-localization -review-assets
.github/instructions/** review-agent-guidance -review-implementation -review-test -review-documentation -review-generated -review-localization -review-assets
agents/** review-agent-guidance -review-implementation -review-test -review-documentation -review-generated -review-localization -review-assets
skills/** review-agent-guidance -review-implementation -review-test -review-documentation -review-generated -review-localization -review-assets
prompts/** review-agent-guidance -review-implementation -review-test -review-documentation -review-generated -review-localization -review-assets
instructions/** review-agent-guidance -review-implementation -review-test -review-documentation -review-generated -review-localization -review-assets
workflows/** review-agent-guidance -review-implementation -review-test -review-documentation -review-generated -review-localization -review-assets
.agents/** review-agent-guidance -review-implementation -review-test -review-documentation -review-generated -review-localization -review-assets
.impeccable/** review-agent-guidance -review-implementation -review-test -review-documentation -review-generated -review-localization -review-assets
memory/** review-agent-guidance -review-implementation -review-test -review-documentation -review-generated -review-localization -review-assets
memory.md review-agent-guidance -review-implementation -review-test -review-documentation -review-generated -review-localization -review-assets

# Tests. Last, so a test bundled inside skills/ or agents/ is a test and not
**/__tests__/** review-test -review-implementation -review-documentation -review-generated -review-agent-guidance -review-localization -review-assets
tests/** review-test -review-implementation -review-documentation -review-generated -review-agent-guidance -review-localization -review-assets
test/** review-test -review-implementation -review-documentation -review-generated -review-agent-guidance -review-localization -review-assets
__tests__/** review-test -review-implementation -review-documentation -review-generated -review-agent-guidance -review-localization -review-assets
*.test.js review-test -review-implementation -review-documentation -review-generated -review-agent-guidance -review-localization -review-assets
*.test.cjs review-test -review-implementation -review-documentation -review-generated -review-agent-guidance -review-localization -review-assets
*.test.mjs review-test -review-implementation -review-documentation -review-generated -review-agent-guidance -review-localization -review-assets
*.test.ts review-test -review-implementation -review-documentation -review-generated -review-agent-guidance -review-localization -review-assets
*.spec.js review-test -review-implementation -review-documentation -review-generated -review-agent-guidance -review-localization -review-assets
*.spec.ts review-test -review-implementation -review-documentation -review-generated -review-agent-guidance -review-localization -review-assets
*.bats review-test -review-implementation -review-documentation -review-generated -review-agent-guidance -review-localization -review-assets