lychee.toml

verbose = "debug"
no_progress = false
exclude_all_private = false
max_concurrency = 16
max_retries = 3
retry_wait_time = 5
cache_exclude_status = [429]
accept = [200, 403, 429, 502, 503] # 403 is often returned by private repos instead of 404; 429 is rate limiting; 502/503 is transient GitHub unavailability
exclude_path = [
    "crates/utilities/test-utils/contracts/",
    "crates/shared/primitives/contracts/",
    "docs/specs/",
    "README\\.md",
]
exclude = [
    # Internal / placeholder
    'foo.bar',
    'localhost',

    # Shields.io badges that query live GitHub stats (flaky, not real links)
    'https://img.shields.io/github/commit-activity/w/base/base',
    'https://img.shields.io/github/issues-pr-raw/base/base',

    # Sites that block automated HTTP clients / scrapers
    'https://.*etherscan\.io',         # Block explorer; returns 403/429 for bots
    'https://medium\.com',             # Medium blocks crawlers
    'https://.*\.medium\.com',
    'https://twitter\.com',            # Returns 4xx without a browser UA
    'https://www\.man7\.org',          # Linux man pages, blocks crawlers
    'https://www\.linux-mips\.org',
    'https://patreon\.com',
    'https://www\.patreon\.com',
    'https://tidelift\.com',

    # Internal / staging sites not meant for link checking
    'https://rustup\.rs/',
    'https://github\.com/base/node/pkgs/container/node-reth', # GitHub container registry pages return intermittent 502s
    'https://base\.mirror\.xyz/',
    'https://specs\.base\.org',

    # High-volume npm registry URLs from lockfiles (belt-and-suspenders alongside
    # the package-lock.json exclude_path; registry URLs are not meaningful doc links)
    'https://registry\.npmjs\.org',
    'https://opencollective\.com',     # Funding pages; rate-limits and redirects heavily
]