feat(import-resilience L3): subprocess-isolated probes for video + archive

Layer 3 — prevent the hard worker crash rather than just recovering from
it. The realistic process-crash vectors (operator's observed slow/heavy
tasks) are video decode and archive extraction; images decode in-process
and Pillow raises-and-skips cleanly, and a subprocess per image would
wreck deep-scan throughput, so images are intentionally not probed.

New backend/app/utils/safe_probe.py (leaf module, lazy heavy imports so
the spawned child stays light):

- probe_video(path): validates the container + first video stream via
  ffprobe (a separate binary — a decoder crash kills only ffprobe, not
  the worker). Returns width/height, which the importer didn't capture
  for videos before. crashed=True only on ffprobe timeout.
- probe_archive(path): an uncompressed-size bomb guard
  (MAX_ARCHIVE_UNCOMPRESSED_BYTES = 4 GiB) plus the format integrity
  test (zipfile.testzip / rarfile.testrar / py7zr.test) run in a
  spawned child process. A decompression-bomb OOM or native-lib
  segfault on a malformed archive shows up as a non-zero child exit
  code → crashed=True, never a dead worker.

ProbeResult.crashed distinguishes a HARD failure (subprocess killed /
timed out — the poison-pill signature → caller returns terminal
'failed') from a CLEAN rejection (corrupt-but-handled, bomb cap,
integrity mismatch → caller's choice of skipped/attached).

Wired:
- importer._import_media video branch: probe_video before the pipeline;
  crash → failed, clean reject → invalid_image skip, ok → capture dims.
- importer._import_archive: probe_archive before extract_archive; crash
  → failed, clean reject → still preserve the archive as a
  PostAttachment (matches extract_archive's fail-soft contract).
- ml.tag_and_embed video branch: probe_video before sampling 10 frames,
  so a corrupt video is rejected (status='bad_video') instead of
  crashing the ml-worker on frame decode.

Tests (test_safe_probe.py): valid/corrupt zip via probe_archive, direct
_inspect_archive size+integrity, in-process _archive_probe_target bomb
guard (monkeypatch can't reach a spawned child, so the target is called
directly), and a non-video → ok=False that's robust to ffprobe presence
in CI.

backend/app/services/importer.py

@@ -30,6 +30,7 @@ from ..models import (
     PostAttachment,
     Source,
 )
+from ..utils import safe_probe
 from ..utils.paths import derive_subdir, derive_top_level_artist, hash_suffixed_name
 from ..utils.phash import compute_phash, find_similar
 from ..utils.sidecar import find_sidecar, parse_sidecar
@@ -407,6 +408,29 @@ class Importer:
         return ImportResult(status="attached")
 
     def _import_archive(self, source: Path) -> ImportResult:
+        # Layer-3 isolation: bomb-size guard + integrity test in a
+        # spawned child BEFORE extracting in this process. A
+        # decompression bomb or a native-lib crash on a malformed
+        # archive is contained to the child; we reject the file cleanly
+        # instead of OOMing/segfaulting the import worker. extract_archive
+        # is already fail-soft for plain exceptions, so this only adds
+        # the hard-crash protection.
+        probe = safe_probe.probe_archive(source)
+        if not probe.ok:
+            if probe.crashed:
+                return ImportResult(
+                    status="failed",
+                    error=f"archive probe crashed/timed out: {probe.reason}",
+                )
+            # Clean rejection (bomb cap exceeded, integrity mismatch):
+            # still preserve the archive file itself as an attachment so
+            # nothing silently vanishes, matching extract_archive's
+            # fail-soft contract.
+            artist = self._resolve_artist(source)
+            post = self._post_for_sidecar(source, artist)
+            self._capture_attachment(source, post=post, artist=artist, resolved=True)
+            return ImportResult(status="attached")
+
         artist = self._resolve_artist(source)
         post = self._post_for_sidecar(source, artist)
         member_ids: list[int] = []
@@ -446,7 +470,25 @@ class Importer:
         # Compute file dimensions (images only) and apply filters.
         width = height = None
         has_alpha = False
-        if not is_video(source):
+        if is_video(source):
+            # Layer-3 isolation: validate the container via ffprobe (a
+            # separate process) before the rest of the pipeline touches
+            # it. A corrupt video that would crash a decoder is rejected
+            # cleanly here, and we capture width/height for free (the
+            # importer didn't previously record video dimensions).
+            probe = safe_probe.probe_video(source)
+            if not probe.ok:
+                if probe.crashed:
+                    return ImportResult(
+                        status="failed",
+                        error=f"video probe crashed/timed out: {probe.reason}",
+                    )
+                return ImportResult(
+                    status="skipped", skip_reason=SkipReason.invalid_image,
+                    error=probe.reason,
+                )
+            width, height = probe.width, probe.height
+        else:
             try:
                 with Image.open(source) as im:
                     im.verify()