What this tests

headersparsing

Point your client at this specimen and find out — before prod does — whether it reads the headers, or just the status code.

Anatomy

The Content-Type says charset=utf-16. The body is UTF-8. The label and the bytes are in different encodings and the client gets to decide which to believe — trust the header and decode UTF-8 as UTF-16, which produces an elegant pile of CJK-adjacent garbage, or sniff the bytes and ignore the declared metadata. Different from FD-2026-014's Latin-1 gambit: there the header was right about UTF-8 and the bytes were wrong. Here the header is the lie and the bytes are honest — the manifest was stamped wrong at the dock, not the cargo. Both are "encoding mismatch" the way "fire" and "flood" are both "water damage". Clients that sniff survive. Clients that obey the declared charset decode a response that looks like it was written by a very enthusiastic Unicode accident.

Replay it

curl -i https://faultydata.com/r/FD-2026-086

// response appears here

Signature

{Status:200 Headers:map[Content-Type:application/json; charset=utf-16] Body:{"id":12345,"name":"ada","note":"the bytes are honest UTF-8. the header is lying."} Corrupt: Delay:<nil> Flaky:<nil> Sequence:<nil> Drip:<nil> StallAfter:0 Hang:false Reset:false Raw: Phases:[]}

Keep digging

FD-2026-019 — JSON body, text/plain label FD-2026-041 — Retry-After as HTTP-date FD-2026-100 — Conflicting duplicate Content-Length headers FD-2026-036 — Content-Encoding lie (gzip header, plain body) FD-2026-039 — XML where JSON was promised FD-2026-052 — JSONP callback wrapper