# Allowlist for scripts/check-website-urls.sh (substring match against the URL).
# Use ONLY for URLs that are valid but 4xx an automated bare probe (API/CDN
# bases, OIDC endpoints) or are intentionally illustrative placeholders in prose.
#
# localhost / 127.0.0.1 / example.com / example.org and the w3.org SVG XML
# namespace are skipped intrinsically and do NOT need entries here.
#
# 401/403/429 (auth walls, bot blocks, rate limits — e.g. npmjs.com, openai.com)
# are already classified as "blocked" advisories, NOT dead links, so they do
# NOT need entries here either.

# Google Fonts CSS/font CDN — 404s a bare base probe, valid with a font query.
fonts.googleapis.com
fonts.gstatic.com

# Gemini OpenAI-compat API base — 404s without a path/auth; cited as a base URL.
generativelanguage.googleapis.com

# GitHub Actions OIDC token endpoint — 404s a bare GET; cited in CI/OIDC docs.
token.actions.githubusercontent.com

# Illustrative placeholder org in SKILLCARD registry examples (not a real repo).
github.com/my-org/

# Real repo (confirmed by the maintainer); GitHub 404s blob deep-links to bot
# probes, and some referenced skill files (e.g. dentist/SKILL.md) are illustrative.
github.com/Atmosphere/atmosphere-skills

# Documented LLM/provider API base endpoints — 404/405 a bare browser probe
# (POST + auth only), cited as base URLs in prose. Not link rot.
api.anthropic.com
api.openai.com
api.cohere.com
api.telegram.org
