The self-referential loop — where discussing domain fragments in chat created new articles containing those fragments — led to a behavioral rule: never type provider-domain strings in chat. Combined with the emit-time redaction filter, this creates defense-in-depth privacy protection. The bug that exposed the vulnerability also produced the safeguard.