v4.0.0: codec 2.60.4 security release, --pq-box sealed-box mode, F-16 fix

Major release. Highlights:

- Codec: vendored VaptVupt codec moves to canonical 2.60.4 security
  release. Fixes a high-severity OOB heap write in the AVX2 decode fast
  path (reachable on a valid stream sized to exactly content_size, both
  tail variants). Brings CBMC-formally-verified BCJ filters with
  automatic ELF/PE/Mach-O detection. Compressed output stays
  byte-identical (ratio gate Δ 0.00%); wire format unchanged at v1.6.
- New --pq-box sealed-box recipient mode (vendored libpqvaptvupt 0.6.0):
  ML-KEM-768 + X25519 combined via HKDF-SHA256 with domain separation,
  AES-256-CTR + HMAC-SHA256 EtM. Legacy --pq and --pq-sdk stay readable.
- F-16: discloses and fixes a pre-existing data-loss defect in the
  <= 3.8.0 in-tree BCJ encoder. Full back-compat matrix decodes
  byte-exact under 4.0.0; every readable pre-4.0 archive remains readable.

Repository hygiene:
- Sync full 4.0.0 source tree (codec, crypto, SDK, GUI, packaging, tests).
- Remove internal scratch files (PROMPT.md, FORMAL_AUDIT_PROMPT.md)
  and superseded version-specific docs (INTEGRATION_PROTOCOL_2.60.4.md,
  docs/FINDINGS-2.x.md) and a stray test binary.
- Refresh README download/install section to real 4.0.0 release assets;
  bump version badge to 4.0.0.
- Add .gitignore for build outputs (keeps vendored prebuilt libraries).

include/vv_bcj.h

@@ -0,0 +1,57 @@
+/*
+ * SPDX-License-Identifier: GPL-3.0-or-later
+ * VaptVupt — BCJ branch filters (see src/vv_bcj.c).
+ *
+ * Reversible, architecture-specific branch converters that improve the
+ * compression of machine code by turning relative call targets into an
+ * absolute form. Each is an exact bijection on arbitrary input, so a file
+ * filtered with the wrong architecture (or no machine code at all) still
+ * round-trips byte-for-byte.
+ */
+#ifndef VV_BCJ_H
+#define VV_BCJ_H
+
+#include <stddef.h>
+#include <stdint.h>
+
+/*
+ * x86 / x86-64 BCJ. Converts near CALL (0xE8) and JMP (0xE9) relative
+ * displacements to/from absolute. encoding != 0 = forward (compress-side),
+ * 0 = inverse (decode-side). `ip` is the stream offset of byte 0 (use 0 for
+ * whole-buffer transforms). Returns the prefix length that may have been
+ * modified. vv_bcj_x86(b,n,0,0) undoes vv_bcj_x86(b,n,0,1).
+ */
+size_t vv_bcj_x86(uint8_t *data, size_t size, uint32_t ip, int encoding);
+
+/*
+ * AArch64 (ARM64) BL + ADRP filter. Converts BL (call) 26-bit relative word
+ * offsets and ADRP (PC-relative page address) 21-bit page offsets to/from an
+ * absolute form, each modulo its immediate width. Same calling convention as
+ * vv_bcj_x86. Only BL (opcode 100101) and ADRP (1xx10000) are touched;
+ * opcode and register bits are preserved, so the transform is an exact
+ * bijection on arbitrary input. vv_bcj_arm64(b,n,0,0) undoes
+ * vv_bcj_arm64(b,n,0,1).
+ */
+size_t vv_bcj_arm64(uint8_t *data, size_t size, uint32_t ip, int encoding);
+
+/* Which branch filter best fits a buffer, by sniffing an executable header. */
+typedef enum {
+    VV_FILTER_NONE  = 0,
+    VV_FILTER_X86   = 1,
+    VV_FILTER_ARM64 = 2
+} vv_filter_kind_t;
+
+/*
+ * Inspect the first bytes of `data` for an ELF, PE (MZ/PE), or Mach-O header
+ * and return the BCJ filter that matches its machine type:
+ *   - x86 / x86-64 (and 32-bit x86)        -> VV_FILTER_X86
+ *   - AArch64 (ARM64)                       -> VV_FILTER_ARM64
+ *   - anything else, or no recognised header-> VV_FILTER_NONE
+ * Fully bounds-checked: safe on truncated or arbitrary input. Detection
+ * errors are never correctness bugs — a missed match just means no filter,
+ * and a spurious match still round-trips (the filters are bijections), it
+ * merely may not improve the ratio.
+ */
+vv_filter_kind_t vv_bcj_detect(const uint8_t *data, size_t size);
+
+#endif /* VV_BCJ_H */