STATUS.md drift: gamut-dsp lists a shipped module as deferred, gamut-avif contradicts itself

[justin13888]

Found while correcting the README crates table for #425 and outside PR #526's boundary, which is
frozen to README.md and the guard that reads it.

#425's working premise was that a crate's STATUS.md is the authority and the README is what had
drifted. That premise has a hole: two STATUS.md files are themselves wrong, and a row derived
faithfully from them is wrong as a result. PR #526 has adopted the rule that where a STATUS.md
contradicts the crate's lib.rs, lib.rs wins — a README row has to be true of the crate, not
merely faithful to a stale file. The stale files themselves are left for this issue.

1. gamut-dsp/STATUS.md lists a module the crate ships

crates/gamut-dsp/STATUS.md:67, under Deferred / tracked follow-ups:

• jpeg module — the ITU-T T.81 8×8 DCT (and friends) for gamut-tiff's JPEG-in-TIFF …

But the module exists and is public:

• crates/gamut-dsp/src/lib.rs:43 — pub mod jpeg;
• crates/gamut-dsp/src/lib.rs:6-7 — the crate docs list jpeg as one of the shipped spec families
• crates/gamut-dsp/src/jpeg/dct.rs:112 — pub fn fdct8x8(block: &mut [i32; 64])
• crates/gamut-dsp/src/jpeg/dct.rs:163 — pub fn idct8x8(block: &mut [i32; 64])

Line 15 of the same file still speaks of jpeg as "a future … module". The consequence was
concrete: #425 narrowed the README's gamut-dsp row to AV1-only wording on this file's authority,
and an independent review caught it.

What is genuinely still deferred is the consumer — JPEG-in-TIFF in gamut-tiff — not the kernel.
The entry should move out of the deferred list and say so.

2. gamut-avif/STATUS.md contradicts itself

crates/gamut-avif/STATUS.md:21-24 is contradicted by later lines in the same file. This overlaps
PR #506's territory, so it is named here for completeness rather than claimed: whoever lands #506
should reconcile it, and if #506 already does, this half can be closed as covered.

Scope

• Move gamut-dsp's jpeg module out of "Deferred / tracked follow-ups" and restate what is
  actually deferred (the gamut-tiff JPEG-in-TIFF consumer). Fix line 15's "a future jpeg …".
• Reconcile gamut-avif/STATUS.md:21-24 with the rest of that file, coordinating with feat(avif): reserve and write a C2PA manifest store, and locate one on read #506.
• Consider whether a STATUS.md claiming a module is deferred while lib.rs declares it pub is
  mechanically checkable, in the way mise run check-readme-crates made the README's crates table
  checkable. A "deferred module that is pub mod" is a grep, not a judgement call.

Acceptance

No STATUS.md lists as deferred a module its crate's lib.rs declares pub, and
gamut-avif/STATUS.md says one thing about its own scope.