Well-known files that misbehave.
Crawlers, scanners, and AI agents discover sites by fetching a small
set of canonical-path files: robots.txt,
llms.txt, sitemap.xml, ads.txt,
/.well-known/security.txt, and friends. Most tools treat
these as trusted metadata.
bots.catastrophic.io serves chaotic versions of each at
the exact path real tools fetch. Point your crawler, security scanner,
agent framework, or DSP here and observe how it handles contradictions,
prompt injection, expired-required-fields, and other malformed metadata.
Every response carries an X-Chaos-*-Mode header so
monitoring clients can verify which flavor of chaos they received.
Each path also has a well-formed counterpart at the same URL on
not.catastrophic.io.
4 flavors of broken metadata
Each category groups related canonical-path files. Modes within each file are selected by ?mode= and reflected in response headers.
App linking
Mobile and PWA discovery files that mishandle the strict format requirements iOS, Android, and browsers enforce.
Conflicting discovery
Three sibling well-known files that each claim a different authoritative host. Tests whether agents reconcile contradictions or trust the first source they see.
Crawler files
Chaotic versions of the well-known files that crawlers, scanners, and AI agents fetch by canonical path.
Federated identity
Discovery files used by federated networks (Fediverse, AT Protocol) that misadvertise software identity or DID ownership.