Merge pull request 'Ship: system-tag refactor + soft WIP tier + Explore reach (#1464, #1474, #1476)' (#223) from dev into main

alembic/versions/0086_process_auto_apply_settings.py

@@ -0,0 +1,61 @@
+"""process auto-apply settings + review mode (#1464) — system-tag refactor
+
+The system-tag behavior refactor gives `wip` / `editor screenshot` (the PROCESS
+group) their own provisional auto-apply, parallel to the presentation (chrome)
+sweep. MLSettings gains three knobs: enabled (OFF by default — a new whole-library
+auto-tagger is opt-in), the flat apply threshold, and the ring-loud conflict
+threshold. presentation_review gains a `mode` column so one review surface serves
+both chrome and process flags (existing rows backfill 'chrome'). server_defaults
+so the existing rows fill cleanly.
+
+Revision ID: 0086
+Revises: 0085
+Create Date: 2026-07-13
+"""
+from typing import Sequence, Union
+
+import sqlalchemy as sa
+from alembic import op
+
+revision: str = "0086"
+down_revision: Union[str, None] = "0085"
+branch_labels: Union[str, Sequence[str], None] = None
+depends_on: Union[str, Sequence[str], None] = None
+
+
+def upgrade() -> None:
+    op.add_column(
+        "ml_settings",
+        sa.Column(
+            "process_auto_apply_enabled", sa.Boolean(), nullable=False,
+            server_default=sa.text("false"),
+        ),
+    )
+    op.add_column(
+        "ml_settings",
+        sa.Column(
+            "process_auto_apply_threshold", sa.Float(), nullable=False,
+            server_default="0.90",
+        ),
+    )
+    op.add_column(
+        "ml_settings",
+        sa.Column(
+            "process_conflict_threshold", sa.Float(), nullable=False,
+            server_default="0.50",
+        ),
+    )
+    op.add_column(
+        "presentation_review",
+        sa.Column(
+            "mode", sa.String(16), nullable=False,
+            server_default="chrome",
+        ),
+    )
+
+
+def downgrade() -> None:
+    op.drop_column("presentation_review", "mode")
+    op.drop_column("ml_settings", "process_conflict_threshold")
+    op.drop_column("ml_settings", "process_auto_apply_threshold")
+    op.drop_column("ml_settings", "process_auto_apply_enabled")

alembic/versions/0087_wip_soft_title_tagging.py

@@ -0,0 +1,33 @@
+"""soft WIP title tier toggle (#1474) — ImportSettings.wip_soft_title_tagging_enabled
+
+The soft tier also tags sketch/doodle/scribble titles, but with a provisional source
+that never trains the head. OFF by default (a lower-precision tier is opt-in).
+server_default so the existing singleton row (id=1) fills cleanly.
+
+Revision ID: 0087
+Revises: 0086
+Create Date: 2026-07-13
+"""
+from typing import Sequence, Union
+
+import sqlalchemy as sa
+from alembic import op
+
+revision: str = "0087"
+down_revision: Union[str, None] = "0086"
+branch_labels: Union[str, Sequence[str], None] = None
+depends_on: Union[str, Sequence[str], None] = None
+
+
+def upgrade() -> None:
+    op.add_column(
+        "import_settings",
+        sa.Column(
+            "wip_soft_title_tagging_enabled", sa.Boolean(), nullable=False,
+            server_default=sa.text("false"),
+        ),
+    )
+
+
+def downgrade() -> None:
+    op.drop_column("import_settings", "wip_soft_title_tagging_enabled")

backend/app/api/gallery.py

@@ -148,6 +148,17 @@ async def similar():
     # Explore passes exclude_wip=1 to also drop work-in-progress from the
     # rabbit-hole; the gallery's own "similar" button omits it (keeps wip, #1274).
     exclude_wip = request.args.get("exclude_wip") in ("1", "true", "True")
+    # Explore reach (#1476): 0 = nearest (gallery default), →1 reaches into farther
+    # distance bands so the walk can escape a dense cluster. exclude_ids = the
+    # breadcrumb, so already-walked images aren't re-served as neighbours.
+    try:
+        reach = max(0.0, min(1.0, float(request.args.get("reach", "0"))))
+    except ValueError:
+        reach = 0.0
+    exclude_ids = [
+        int(x) for x in request.args.get("exclude_ids", "").split(",")
+        if x.strip().isdigit()
+    ] or None
     # post_id is the exclusive post-detail view — not a similarity scope.
     # include_hidden is a gallery-browse flag; similar() has its OWN presentation
     # exclusion (a similarity-quality concern, #1274), so drop it here (#141).
@@ -158,7 +169,8 @@ async def similar():
         svc = GalleryService(session)
         try:
             images = await svc.similar(
-                image_id=similar_to, limit=limit, exclude_wip=exclude_wip, **scope)
+                image_id=similar_to, limit=limit, exclude_wip=exclude_wip,
+                reach=reach, exclude_ids=exclude_ids, **scope)
         except ValueError as exc:
             return jsonify({"error": str(exc)}), 400
     if images is None:
@@ -236,8 +248,10 @@ async def jump():
 #    content", surfaced in the gallery's Show-hidden review strip. -----------
 @gallery_bp.route("/hidden-review", methods=["GET"])
 async def hidden_review():
-    """Unresolved presentation auto-hide flags, most-concerning first (highest
-    content score) — for the gallery's Hidden-view review strip."""
+    """Unresolved system-tag auto-apply review flags (chrome + process, #1464),
+    most-concerning first (highest content score) — for the review strip. `mode`
+    tells the client whether the flagged tag hid the image ('chrome') or left it
+    visible ('process'), which decides the resolve labels (un-hide vs remove-tag)."""
     ptag = aliased(Tag)
     ctag = aliased(Tag)
     async with get_session() as session:
@@ -247,6 +261,7 @@ async def hidden_review():
                 PresentationReview.tag_id,
                 PresentationReview.conflict_tag_id,
                 PresentationReview.conflict_score,
+                PresentationReview.mode,
                 ImageRecord.path, ImageRecord.thumbnail_path,
                 ImageRecord.sha256, ImageRecord.mime,
                 ptag.name.label("tag_name"),
@@ -266,6 +281,7 @@ async def hidden_review():
             "conflict_tag_id": r.conflict_tag_id,
             "conflict_name": r.conflict_name,
             "conflict_score": r.conflict_score,
+            "mode": r.mode,
             "thumbnail_url": thumbnail_url(r.thumbnail_path, r.sha256, r.mime),
             "image_url": image_url(r.path),
         }

backend/app/api/ml_admin.py

@@ -42,6 +42,9 @@ _EDITABLE = (
     "presentation_auto_apply_enabled",
     "presentation_auto_apply_threshold",
     "presentation_conflict_threshold",
+    "process_auto_apply_enabled",
+    "process_auto_apply_threshold",
+    "process_conflict_threshold",
     "embedder_model_name",
     "embedder_model_version",
     *_DETECTOR_FIELDS,
@@ -102,6 +105,9 @@ async def get_settings():
             "presentation_auto_apply_enabled": s.presentation_auto_apply_enabled,
             "presentation_auto_apply_threshold": s.presentation_auto_apply_threshold,
             "presentation_conflict_threshold": s.presentation_conflict_threshold,
+            "process_auto_apply_enabled": s.process_auto_apply_enabled,
+            "process_auto_apply_threshold": s.process_auto_apply_threshold,
+            "process_conflict_threshold": s.process_conflict_threshold,
             "embedder_model_name": s.embedder_model_name,
             **{f: getattr(s, f) for f in _DETECTOR_FIELDS},
         }
@@ -162,6 +168,12 @@ def _validate(p: dict) -> str | None:
         return "presentation_auto_apply_threshold must be between 0.5 and 0.999"
     if not (0.0 <= float(p["presentation_conflict_threshold"]) <= 1.0):
         return "presentation_conflict_threshold must be between 0 and 1"
+    # Process auto-apply (#1464). wip/editor stay VISIBLE so a false apply is
+    # low-harm (excludes-from-training + a review flag), but keep the same bar.
+    if not (0.5 <= float(p["process_auto_apply_threshold"]) <= 0.999):
+        return "process_auto_apply_threshold must be between 0.5 and 0.999"
+    if not (0.0 <= float(p["process_conflict_threshold"]) <= 1.0):
+        return "process_conflict_threshold must be between 0 and 1"
     # Embedder model swap (#1190): both must be non-empty. Changing them means a
     # different embedding space — the operator must re-embed + retrain after.
     for key in ("embedder_model_name", "embedder_model_version"):

backend/app/api/settings.py

@@ -49,6 +49,7 @@ _EDITABLE_FIELDS = (
     "translation_target_lang",
     "translation_min_confidence",
     "wip_title_tagging_enabled",
+    "wip_soft_title_tagging_enabled",
 )
 
 # Per-host external-download toggles — all plain booleans, validated uniformly.
@@ -91,6 +92,7 @@ async def get_import_settings():
         "translation_target_lang": row.translation_target_lang,
         "translation_min_confidence": row.translation_min_confidence,
         "wip_title_tagging_enabled": row.wip_title_tagging_enabled,
+        "wip_soft_title_tagging_enabled": row.wip_soft_title_tagging_enabled,
     })
 
 
@@ -179,6 +181,12 @@ async def update_import_settings():
         return jsonify(
             {"error": "wip_title_tagging_enabled must be a boolean"}
         ), 400
+    if "wip_soft_title_tagging_enabled" in body and not isinstance(
+        body["wip_soft_title_tagging_enabled"], bool
+    ):
+        return jsonify(
+            {"error": "wip_soft_title_tagging_enabled must be a boolean"}
+        ), 400
 
     async with get_session() as session:
         row = await ImportSettings.load(session)

backend/app/celery_app.py

@@ -171,9 +171,19 @@ def make_celery() -> Celery:
             },
             "presentation-auto-apply-daily": {
                 "task": "backend.app.tasks.ml.scheduled_presentation_auto_apply",
-                "schedule": 86400.0,  # auto-hide banner/editor chrome (#141);
+                "schedule": 86400.0,  # auto-hide banner chrome (#141);
                 # no-op unless presentation_auto_apply_enabled
             },
+            "process-auto-apply-daily": {
+                "task": "backend.app.tasks.ml.scheduled_process_auto_apply",
+                "schedule": 86400.0,  # auto-tag wip/editor process art (#1464);
+                # no-op unless process_auto_apply_enabled (opt-in)
+            },
+            "soft-wip-conflict-audit-daily": {
+                "task": "backend.app.tasks.ml.scheduled_soft_wip_conflict_audit",
+                "schedule": 86400.0,  # flag ring-loud soft-WIP (sketch/doodle) tags
+                # for review (#1474); no-op with no content heads
+            },
             "prune-presentation-reviews-daily": {
                 "task": "backend.app.tasks.ml.prune_presentation_reviews",
                 "schedule": 86400.0,  # retention: drop resolved review flags >30d

backend/app/models/import_settings.py

@@ -126,6 +126,13 @@ class ImportSettings(Base):
     wip_title_tagging_enabled: Mapped[bool] = mapped_column(
         Boolean, nullable=False, default=True, server_default="true",
     )
+    # Soft WIP title tier (#1474): also tag sketch/doodle/scribble titles, but with
+    # a PROVISIONAL source (`wip_title_soft`) that never trains the head, since these
+    # are lower-precision (a finished "sketch" isn't WIP). OFF by default — a lower-
+    # precision tier is opt-in (the ring-loud audit surfaces false positives).
+    wip_soft_title_tagging_enabled: Mapped[bool] = mapped_column(
+        Boolean, nullable=False, default=False, server_default="false",
+    )
 
     @classmethod
     async def load(cls, session) -> ImportSettings:

backend/app/models/ml_settings.py

@@ -85,12 +85,14 @@ class MLSettings(Base):
         Float, nullable=False, default=0.95
     )
     # -- Presentation chrome auto-hide (#141) -------------------------------
-    # banner / editor screenshot auto-apply on the sweep with their OWN flat
-    # threshold (decoupled from content-head graduation). Hiding is consequential
-    # so it runs HIGH. `wip` is never auto-applied. When an image would be
-    # auto-hidden but ALSO scores >= presentation_conflict_threshold on a content
-    # head, it's still hidden but flagged for review (PresentationReview) instead
-    # of buried silently. ON by default (opt-out); every auto-tag is reversible.
+    # `banner` (chrome — clusters on UI, not content) auto-applies on the sweep
+    # with its OWN flat threshold (decoupled from content-head graduation) and is
+    # HIDDEN from the gallery. Hiding is consequential so it runs HIGH. When an
+    # image would be auto-hidden but ALSO scores >= presentation_conflict_threshold
+    # on a content head, it's still hidden but flagged for review
+    # (PresentationReview, mode='chrome') instead of buried silently. ON by default
+    # (opt-out); every auto-tag is reversible. NOTE (#1464): `wip` + `editor
+    # screenshot` are no longer chrome — they went to the PROCESS path below.
     presentation_auto_apply_enabled: Mapped[bool] = mapped_column(
         Boolean, nullable=False, default=True
     )
@@ -100,6 +102,26 @@ class MLSettings(Base):
     presentation_conflict_threshold: Mapped[float] = mapped_column(
         Float, nullable=False, default=0.50
     )
+    # -- Process auto-apply (#1464) ----------------------------------------
+    # `wip` / `editor screenshot` are PROCESS art — unfinished pieces + program
+    # screenshots that must stay OUT of head/CCIP training but, unlike chrome,
+    # remain VISIBLE in the gallery (operator 2026-07-12). They auto-apply on the
+    # sweep with their OWN flat threshold and a PROVISIONAL source (`process_auto`,
+    # in training_data._AUTO_SOURCES) so the head NEVER trains on its own output —
+    # it learns only from title (`wip_title`) + manual labels, which breaks the
+    # runaway loop. When a process tag would be applied but the image ALSO scores
+    # >= process_conflict_threshold on a content head, it's flagged for review
+    # (PresentationReview, mode='process') rather than silently marked. OFF by
+    # default — a new whole-library auto-tagger is opt-in; every auto-tag reversible.
+    process_auto_apply_enabled: Mapped[bool] = mapped_column(
+        Boolean, nullable=False, default=False
+    )
+    process_auto_apply_threshold: Mapped[float] = mapped_column(
+        Float, nullable=False, default=0.90
+    )
+    process_conflict_threshold: Mapped[float] = mapped_column(
+        Float, nullable=False, default=0.50
+    )
     # Default = SigLIP 2 (so400m, 512px) for new installs (migration 0069);
     # existing libraries keep their stored value until the operator re-embeds.
     embedder_model_version: Mapped[str] = mapped_column(

backend/app/models/presentation_review.py

@@ -1,15 +1,17 @@
-"""PresentationReview — an auto-hidden presentation tag that ALSO looked like
-real content, flagged for operator review (milestone 141).
+"""PresentationReview — a system-tag the auto-apply sweep applied that ALSO looked
+like real content, flagged for operator review (milestone 141 + #1464).
 
-When the auto-apply sweep hides an image as chrome (banner / editor screenshot)
-but the image ALSO scores highly on a content head, it still hides it but records
-this row so the Hidden view can surface it ("⚠ also looks like <conflict tag>")
-for a keep-hidden / un-hide decision. Resolved rows are pruned by retention.
+When a sweep applies a system tag but the image ALSO scores highly on a content
+head, it still applies the tag but records this row so a review strip can surface
+it ("⚠ also looks like <conflict tag>"). Two modes (#1464): 'chrome' (banner —
+image is HIDDEN, review is keep-hidden / un-hide) and 'process' (wip / editor
+screenshot — image stays VISIBLE, review is confirm / remove-tag). Resolved rows
+are pruned by retention.
 """
 
 from datetime import datetime
 
-from sqlalchemy import DateTime, Float, ForeignKey, func
+from sqlalchemy import DateTime, Float, ForeignKey, String, func
 from sqlalchemy.orm import Mapped, mapped_column
 
 from .base import Base
@@ -31,6 +33,12 @@ class PresentationReview(Base):
         ForeignKey("tag.id", ondelete="SET NULL"), nullable=True
     )
     conflict_score: Mapped[float] = mapped_column(Float, nullable=False)
+    # Which sweep flagged this (#1464): 'chrome' (banner, hidden) or 'process'
+    # (wip / editor screenshot, shown). Drives which review strip surfaces it and
+    # what "resolve" means (un-hide vs remove-tag). Existing rows backfill 'chrome'.
+    mode: Mapped[str] = mapped_column(
+        String(16), nullable=False, default="chrome", server_default="chrome"
+    )
     created_at: Mapped[datetime] = mapped_column(
         DateTime(timezone=True), nullable=False, server_default=func.now()
     )

backend/app/models/tag.py

@@ -43,14 +43,19 @@ class TagKind(StrEnum):
     # to keep historic tag rows queryable.
 
 
-# The seeded system tags (migration 0075). PRESENTATION tags additionally
-# hide from whole-image similarity results — they cluster on UI chrome, not
-# content. `wip` is real art: only the training pipelines exclude it.
+# The seeded system tags (migration 0075). Two behavior groups (#1464):
+#   CHROME (banner): clusters on UI chrome, not content → HIDDEN from the default
+#     gallery + from similarity; auto-applied via the sweep's chrome mode.
+#   PROCESS (wip, editor screenshot): real-but-unfinished art / program screenshots
+#     → SHOWN in the gallery (operator 2026-07-12), but excluded from the Explore
+#     rabbit-hole; auto-applied via the sweep's process mode (provisional source,
+#     ring-loud review guard).
+# ALL three are excluded from OTHER concepts' head/CCIP training (training-hygiene,
+# keyed on is_system); a system tag's OWN head trains on them — that's what makes
+# auto-flagging work.
 SYSTEM_TAG_NAMES = ("wip", "banner", "editor screenshot")
-PRESENTATION_SYSTEM_TAGS = ("banner", "editor screenshot")
-# `wip` marks real-but-unfinished art. It's kept in the gallery's own "similar"
-# results (#1274), but the Explore rabbit-hole opts to hide it (exclude_wip) so a
-# browse doesn't keep surfacing work-in-progress (operator, 2026-07-08).
+CHROME_SYSTEM_TAGS = ("banner",)
+PROCESS_SYSTEM_TAGS = ("wip", "editor screenshot")
 WIP_SYSTEM_TAG = "wip"
 
 image_tag = Table(

backend/app/services/gallery_service.py

@@ -31,7 +31,7 @@ from ..models import (
     Tag,
     TagPositiveConfirmation,
 )
-from ..models.tag import PRESENTATION_SYSTEM_TAGS, WIP_SYSTEM_TAG, image_tag
+from ..models.tag import CHROME_SYSTEM_TAGS, PROCESS_SYSTEM_TAGS, image_tag
 from .pagination import decode_cursor, encode_cursor
 from .tag_query import (
     fandom_join_alias,
@@ -396,6 +396,25 @@ def _diversify_similar(src, rows, limit, *, dup_threshold=8, lam=0.40):
     return [kept[i] for i in order]
 
 
+def _reach_sample(rows, limit, reach):
+    """From a distance-sorted candidate pool (nearest first), pick a spread of ranks
+    that MIXES near (tag the current cluster) and mid-far (escape it) BEFORE dedup +
+    MMR — the Explore "reach" dial (#1476).
+
+    reach in (0, 1]: the sampled span grows outward from the anchor (0.25→1.0 of the
+    pool), evenly strided from rank 0 so the nearest are still represented. In a
+    dense signature the nearest ranks are near-identical, so reaching farther is the
+    only way to hand MMR genuinely different content — MMR alone can't escape a pool
+    that's already all-near. reach<=0 or a small pool passes through unchanged."""
+    n = len(rows)
+    want = max(limit * 8, 100)
+    if reach <= 0 or n <= want:
+        return rows
+    span = int(min(1.0, 0.25 + 0.75 * reach) * n)
+    idx = sorted({min(int(i * span / want), n - 1) for i in range(want)})
+    return [rows[i] for i in idx]
+
+
 async def _artists_for(session, image_ids: list[int]) -> dict[int, dict]:
     """Map image_id -> {"name","slug"} via the canonical
     image_record.artist_id (FC-2d-vii-c). Bounded by page size."""
@@ -419,16 +438,17 @@ class GalleryService:
     async def _hidden_tag_ids(
         self, include_hidden, tag_ids, tag_or_groups,
     ) -> list[int] | None:
-        """Presentation-chrome tag ids to implicitly exclude from a gallery query,
-        or None. None when the caller asked to include hidden, when the operator
-        is explicitly filtering FOR a presentation tag (they clearly want to see
-        it), or when no presentation tags exist. (milestone 141)"""
+        """Chrome (banner) tag ids to implicitly exclude from a gallery query, or
+        None. None when the caller asked to include hidden, when the operator is
+        explicitly filtering FOR a chrome tag (they clearly want to see it), or when
+        no chrome tags exist. (milestone 141; #1464: editor screenshot is now PROCESS
+        — shown — so only `banner` hides here.)"""
         if include_hidden:
             return None
         rows = await self.session.execute(
             select(Tag.id).where(
                 Tag.is_system.is_(True),
-                Tag.name.in_(PRESENTATION_SYSTEM_TAGS),
+                Tag.name.in_(CHROME_SYSTEM_TAGS),
             )
         )
         pres = [r[0] for r in rows]
@@ -716,6 +736,7 @@ class GalleryService:
         untagged: bool = False, no_artist: bool = False,
         date_from: datetime | None = None, date_to: datetime | None = None,
         exclude_wip: bool = False,
+        reach: float = 0.0, exclude_ids: list[int] | None = None,
     ) -> list[GalleryImage] | None:
         """Visual "more like this": images near `image_id`'s SigLIP embedding
         (pgvector, HNSW-indexed — alembic 0036), then DIVERSIFIED so the result
@@ -744,20 +765,27 @@ class GalleryService:
         # wide pool there's nothing but the near-dupes to choose from. Widened
         # (5×→8×, cap 200→400) so the stronger MMR has genuinely distinct
         # neighbourhoods to reach into for more variance (operator, 2026-07-01).
-        pool_n = min(400, max(limit * 8, 100))
+        # Explore's reach>0 (#1476) widens it a LOT more: in a dense signature the
+        # nearest few hundred are all near-identical, so far-enough candidates only
+        # exist deeper in the ranked pool. _reach_sample then strides across them.
+        if reach > 0:
+            pool_n = min(1000, max(limit * 25, 100))
+        else:
+            pool_n = min(400, max(limit * 8, 100))
         distance = ImageRecord.siglip_embedding.cosine_distance(src.siglip_embedding)
         eff = _effective_date_col()
         stmt = select(ImageRecord, Post.post_date, eff.label("eff"))
         stmt = _outer_join_primary_post(stmt)
-        # Presentation images (banner / editor-screenshot system tags, #128)
-        # cluster on UI chrome rather than content, so near any one of them
-        # they'd fill the grid. Excluded from CANDIDATES only — the anchor
-        # itself may be a banner. `wip` stays surfaced here by default (real art;
-        # only the training pipelines exclude it), but the Explore rabbit-hole
-        # passes exclude_wip to also drop work-in-progress (operator, 2026-07-08).
-        excluded_system_tags = PRESENTATION_SYSTEM_TAGS
+        # Chrome (banner, #128) clusters on UI rather than content, so near any one
+        # of them they'd fill the grid → excluded from CANDIDATES always (the anchor
+        # itself may be a banner). PROCESS art (wip / editor screenshot) stays
+        # surfaced here by default (real content; only the training pipelines exclude
+        # it), but the Explore rabbit-hole passes exclude_wip to also drop the whole
+        # process group so a browse doesn't keep surfacing work-in-progress
+        # (operator, 2026-07-08; #1464 — editor now rides with wip here).
+        excluded_system_tags = CHROME_SYSTEM_TAGS
         if exclude_wip:
-            excluded_system_tags = (*PRESENTATION_SYSTEM_TAGS, WIP_SYSTEM_TAG)
+            excluded_system_tags = (*CHROME_SYSTEM_TAGS, *PROCESS_SYSTEM_TAGS)
         presentation = (
             select(image_tag.c.image_record_id)
             .join(Tag, Tag.id == image_tag.c.tag_id)
@@ -771,6 +799,10 @@ class GalleryService:
             ImageRecord.id != image_id,
             ImageRecord.id.not_in(presentation),
         )
+        # Anti-revisit (#1476): the Explore walk passes its breadcrumb so already-
+        # walked images aren't re-served as neighbours — → can't loop you back in.
+        if exclude_ids:
+            stmt = stmt.where(ImageRecord.id.not_in(exclude_ids))
         stmt = _apply_scope(
             stmt, tag_ids=tag_ids, post_id=None,
             artist_id=artist_id, media_type=media_type,
@@ -780,6 +812,10 @@ class GalleryService:
         )
         stmt = stmt.order_by(distance.asc()).limit(pool_n)
         rows = (await self.session.execute(stmt)).all()
+        # Explore reach: stride across an outward-growing distance span so the pool
+        # handed to MMR spans near→mid-far, not just the tight cluster (#1476).
+        if reach > 0:
+            rows = _reach_sample(rows, limit, reach)
         rows = _diversify_similar(src, rows, limit)
         artists = await _artists_for(self.session, [r[0].id for r in rows])
         return _gallery_images(rows, artists)

backend/app/services/importer.py

@@ -47,7 +47,14 @@ from .attachment_store import AttachmentStore
 from .audits import single_color
 from .link_extract import extract_external_links
 from .thumbnailer import Thumbnailer
-from .wip_title import apply_wip_image_tags, matches_wip_title, resolve_wip_tag_id
+from .wip_title import (
+    WIP_TITLE_SOFT_SOURCE,
+    WIP_TITLE_SOURCE,
+    apply_wip_image_tags,
+    matches_soft_wip_title,
+    matches_wip_title,
+    resolve_wip_tag_id,
+)
 
 log = logging.getLogger(__name__)
 
@@ -999,7 +1006,9 @@ class Importer:
         removal sticks. The existing catalogue is covered separately by the
         operator-triggered backfill sweep. Gated by the settings toggle, and
         best-effort: any failure is logged, never allowed to fail the import."""
-        if not self.settings.wip_title_tagging_enabled:
+        hard_on = self.settings.wip_title_tagging_enabled
+        soft_on = self.settings.wip_soft_title_tagging_enabled
+        if not (hard_on or soft_on):
             return
         if record.primary_post_id is None:
             return
@@ -1007,13 +1016,22 @@ class Importer:
             title = self.session.execute(
                 select(Post.post_title).where(Post.id == record.primary_post_id)
             ).scalar_one_or_none()
-            if not matches_wip_title(title):
+            # HARD tier ("WIP"/"work in progress") wins — higher precision, and it
+            # trains the head; SOFT (sketch/doodle, #1474) is the provisional fallback
+            # that never trains (source wip_title_soft).
+            if hard_on and matches_wip_title(title):
+                source = WIP_TITLE_SOURCE
+            elif soft_on and matches_soft_wip_title(title):
+                source = WIP_TITLE_SOFT_SOURCE
+            else:
                 return
             if self._wip_tag_id is _UNSET:
                 self._wip_tag_id = resolve_wip_tag_id(self.session)
             if self._wip_tag_id is None:
                 return
-            apply_wip_image_tags(self.session, [record.id], self._wip_tag_id)
+            apply_wip_image_tags(
+                self.session, [record.id], self._wip_tag_id, source=source
+            )
         except Exception as exc:  # noqa: BLE001 — a tag must never fail an import
             log.warning(
                 "wip-title auto-tag failed for image %s: %s", record.id, exc

backend/app/services/ml/heads.py

@@ -39,7 +39,7 @@ from ...models import (
     TagPositiveConfirmation,
     TagSuggestionRejection,
 )
-from ...models.tag import PRESENTATION_SYSTEM_TAGS, image_tag
+from ...models.tag import CHROME_SYSTEM_TAGS, PROCESS_SYSTEM_TAGS, image_tag
 from .training_data import (
     _AUTO_SOURCES,
     _auto_apply_point,
@@ -759,18 +759,42 @@ def auto_apply_sweep(
 
 
 _PRESENTATION_SOURCE = "presentation_auto"
+_PROCESS_SOURCE = "process_auto"
+
+# System-tag auto-apply modes (#1464). Both modes run the identical sweep — apply
+# a system tag at a flat threshold with a PROVISIONAL source + a ring-loud review
+# guard — and differ ONLY in which tags, which settings knobs, and which
+# source/review-mode. 'chrome' (banner) is HIDDEN from the gallery; 'process'
+# (wip / editor screenshot) stays VISIBLE (the hide is a gallery-query effect of
+# the tag's group membership, not of this sweep).
+_SWEEP_MODES = {
+    "chrome": {
+        "names": CHROME_SYSTEM_TAGS,
+        "enabled": "presentation_auto_apply_enabled",
+        "threshold": "presentation_auto_apply_threshold",
+        "conflict": "presentation_conflict_threshold",
+        "source": _PRESENTATION_SOURCE,
+    },
+    "process": {
+        "names": PROCESS_SYSTEM_TAGS,
+        "enabled": "process_auto_apply_enabled",
+        "threshold": "process_auto_apply_threshold",
+        "conflict": "process_conflict_threshold",
+        "source": _PROCESS_SOURCE,
+    },
+}
 
 
-def _presentation_heads(session: Session, embedding_version: str):
-    """Trained heads for the presentation chrome tags (banner / editor screenshot).
-    They fire at the FLAT presentation threshold regardless of graduation — a head
-    exists once the operator has labelled enough chrome (head_min_positives)."""
+def _system_tag_heads(session: Session, embedding_version: str, names):
+    """Trained heads for a system-tag group (chrome banner / process wip+editor).
+    They fire at the group's FLAT threshold regardless of graduation — a head
+    exists once the operator has labelled enough (head_min_positives)."""
     return session.execute(
         select(TagHead.tag_id, Tag.name, TagHead.weights, TagHead.bias)
         .join(Tag, Tag.id == TagHead.tag_id)
         .where(TagHead.embedding_version == embedding_version)
         .where(Tag.is_system.is_(True))
-        .where(Tag.name.in_(PRESENTATION_SYSTEM_TAGS))
+        .where(Tag.name.in_(names))
     ).all()
 
 
@@ -802,27 +826,33 @@ def _valued_image_ids(session: Session) -> set[int]:
     return {r[0] for r in rows}
 
 
-def presentation_auto_apply_sweep(session: Session, dry_run: bool = False) -> dict:
-    """Auto-hide presentation chrome (banner / editor screenshot) at the FLAT
-    presentation threshold (#141) — NOT the per-head graduated threshold. Two
-    guards keep it safe: (1) never hide an image carrying a human/confirmed content
-    tag; (2) if an image about to be hidden ALSO scores >= the conflict threshold
-    on a content head, still hide it but flag it (PresentationReview) so the Hidden
-    view surfaces "also looks like <X>" for review. No-op unless
-    presentation_auto_apply_enabled. numpy-only (no sklearn). Returns
-    {n_applied, n_flagged, concepts}."""
+def system_tag_auto_apply_sweep(
+    session: Session, *, mode: str, dry_run: bool = False
+) -> dict:
+    """Auto-apply a system-tag group at its FLAT threshold. mode='chrome' (banner,
+    #141) hides the image; mode='process' (wip / editor screenshot, #1464) keeps it
+    VISIBLE — the ONLY difference is the tag group's gallery membership, not this
+    sweep. Two guards keep it safe: (1) never touch an image carrying a
+    human/confirmed content tag; (2) if the image ALSO scores >= the conflict
+    threshold on a content head, still apply but flag it (PresentationReview,
+    mode=<mode>) so the review strip surfaces "also looks like <X>". The source is
+    PROVISIONAL so the head never trains on its own output. No-op unless the mode's
+    enabled flag is set. numpy-only (no sklearn). Returns {n_applied, n_flagged,
+    concepts}."""
     import numpy as np
     from sqlalchemy.dialects.postgresql import insert as pg_insert
 
+    cfg = _SWEEP_MODES[mode]
     settings = _settings(session)
-    if not dry_run and not settings.presentation_auto_apply_enabled:
+    if not dry_run and not getattr(settings, cfg["enabled"]):
         return {"n_applied": 0, "n_flagged": 0, "concepts": []}
     ver = settings.embedder_model_version
-    pres = _presentation_heads(session, ver)
+    pres = _system_tag_heads(session, ver, cfg["names"])
     if not pres:
         return {"n_applied": 0, "n_flagged": 0, "concepts": []}
-    thr = float(settings.presentation_auto_apply_threshold)
-    conflict_thr = float(settings.presentation_conflict_threshold)
+    thr = float(getattr(settings, cfg["threshold"]))
+    conflict_thr = float(getattr(settings, cfg["conflict"]))
+    source = cfg["source"]
 
     Wp = np.vstack([np.asarray(r.weights, dtype=np.float32) for r in pres])
     bp = np.asarray([r.bias for r in pres], dtype=np.float32)
@@ -884,11 +914,13 @@ def presentation_auto_apply_sweep(session: Session, dry_run: bool = False) -> di
                         pg_insert(image_tag)
                         .values(
                             image_record_id=iid, tag_id=tid,
-                            source=_PRESENTATION_SOURCE,
+                            source=source,
                         )
                         .on_conflict_do_nothing()
                     )
-                # Guard 2: also looks like content → hide but flag for review.
+                # Guard 2: also looks like real content → still apply, but flag it
+                # for the review strip instead of silently marking (chrome hides,
+                # process stays visible — either way the operator gets a heads-up).
                 if Wc is not None and float(max_c[idx]) >= conflict_thr:
                     n_flagged += 1
                     if not dry_run:
@@ -898,6 +930,7 @@ def presentation_auto_apply_sweep(session: Session, dry_run: bool = False) -> di
                                 image_record_id=iid, tag_id=tid,
                                 conflict_tag_id=conf_tag_ids[int(arg_c[idx])],
                                 conflict_score=float(max_c[idx]),
+                                mode=mode,
                             )
                             .on_conflict_do_nothing()
                         )
@@ -914,6 +947,74 @@ def presentation_auto_apply_sweep(session: Session, dry_run: bool = False) -> di
     }
 
 
+def soft_wip_conflict_audit(session: Session, dry_run: bool = False) -> dict:
+    """Ring-loud audit for the SOFT WIP-title cohort (#1474). Images auto-tagged
+    `wip` from a low-precision sketch/doodle title (source='wip_title_soft') that ALSO
+    score >= the process conflict threshold on a content head are probably FINISHED
+    art mis-tagged as process — flag them (PresentationReview, mode='process') so the
+    review strip surfaces them ("also looks like <X>", Keep tag / Remove tag). Does
+    NOT remove the tag; the operator decides. No-op when there are no content heads.
+    numpy-only. Returns {n_scanned, n_flagged}."""
+    import numpy as np
+    from sqlalchemy.dialects.postgresql import insert as pg_insert
+
+    from ..wip_title import WIP_TITLE_SOFT_SOURCE, resolve_wip_tag_id
+
+    settings = _settings(session)
+    ver = settings.embedder_model_version
+    conflict_thr = float(settings.process_conflict_threshold)
+    conf = _conflict_heads(session, ver)
+    wip_id = resolve_wip_tag_id(session)
+    if not conf or wip_id is None:
+        return {"n_scanned": 0, "n_flagged": 0}
+    Wc = np.vstack([np.asarray(r.weights, dtype=np.float32) for r in conf])
+    bc = np.asarray([r.bias for r in conf], dtype=np.float32)
+    conf_tag_ids = [r.tag_id for r in conf]
+
+    soft_ids = [iid for (iid,) in session.execute(
+        select(image_tag.c.image_record_id)
+        .where(image_tag.c.tag_id == wip_id)
+        .where(image_tag.c.source == WIP_TITLE_SOFT_SOURCE)
+    )]
+    # Skip images already flagged for this tag (idempotent re-runs).
+    flagged = {iid for (iid,) in session.execute(
+        select(PresentationReview.image_record_id)
+        .where(PresentationReview.tag_id == wip_id)
+    )}
+    soft_ids = [i for i in soft_ids if i not in flagged]
+
+    n_flagged = 0
+    scanned = 0
+    for start in range(0, len(soft_ids), _AUTO_APPLY_CHUNK):
+        chunk = soft_ids[start:start + _AUTO_APPLY_CHUNK]
+        emb = _load_embeddings(session, chunk)
+        cids = [i for i in chunk if i in emb]
+        if not cids:
+            continue
+        scanned += len(cids)
+        Xn = _l2norm(np.vstack([emb[i] for i in cids]).astype(np.float32), np)
+        cprobs = 1.0 / (1.0 + np.exp(-(Xn @ Wc.T + bc)))
+        max_c = cprobs.max(axis=1)
+        arg_c = cprobs.argmax(axis=1)
+        for k in range(len(cids)):
+            if float(max_c[k]) >= conflict_thr:
+                n_flagged += 1
+                if not dry_run:
+                    session.execute(
+                        pg_insert(PresentationReview)
+                        .values(
+                            image_record_id=cids[k], tag_id=wip_id,
+                            conflict_tag_id=conf_tag_ids[int(arg_c[k])],
+                            conflict_score=float(max_c[k]),
+                            mode="process",
+                        )
+                        .on_conflict_do_nothing()
+                    )
+        if not dry_run:
+            session.commit()
+    return {"n_scanned": scanned, "n_flagged": n_flagged}
+
+
 def retract_auto_applied_heads(session: Session) -> int:
     """Soft auto-apply (milestone 139): re-score every standing source='head_auto'
     tag against its CURRENT head and REMOVE the ones now BELOW the head's

backend/app/services/ml/training_data.py

@@ -29,7 +29,15 @@ from ...models.tag import image_tag
 # a CCIP reference) unless the operator confirms them (milestone 139). Keeping
 # auto-applied predictions out of training is what makes them "soft" — a misfire
 # can't reinforce itself, so the retraction sweep can actually drop it.
-_AUTO_SOURCES = ("head_auto", "ccip_auto", "ml_auto", "presentation_auto")
+# `process_auto` (#1464): wip/editor screenshot applied by the process sweep are
+# ALSO provisional — the head must learn only from title (`wip_title`) + manual
+# labels, never its own auto-applied output, or it would runaway (operator 2026-07-12).
+# `wip_title_soft` (#1474): the soft title tier (sketch/doodle) is LOW-precision, so
+# it's provisional too — a finished piece titled "sketch" must not train the wip head.
+_AUTO_SOURCES = (
+    "head_auto", "ccip_auto", "ml_auto", "presentation_auto", "process_auto",
+    "wip_title_soft",
+)
 
 
 def _hygiene_excluded_ids(session: Session) -> set[int]:

backend/app/services/wip_title.py

@@ -27,7 +27,13 @@ from ..models.tag import WIP_SYSTEM_TAG, Tag, image_tag
 
 # image_tag.source stamped on title-heuristic WIP tags — distinct from the other
 # apply sources so provenance stays legible and a future undo can target only these.
+# HARD tier ("WIP"/"work in progress") is high-precision → trains the wip head.
 WIP_TITLE_SOURCE = "wip_title"
+# SOFT tier (sketch/doodle/scribble, #1474) is LOWER-precision — a finished "sketch"
+# is often not WIP. This source is PROVISIONAL (in training_data._AUTO_SOURCES) so it
+# NEVER trains the wip head; a soft-tagged image that also looks like real content is
+# surfaced by the ring-loud audit for review.
+WIP_TITLE_SOFT_SOURCE = "wip_title_soft"
 
 # A standalone "WIP" / "W.I.P" token, or the phrase "work in progress"
 # (space/underscore/hyphen separated). The letter-boundary lookarounds are what
@@ -39,11 +45,20 @@ _WIP_RE = re.compile(
     re.IGNORECASE,
 )
 
-# Coarse SQL prefilter for the backfill sweep — narrows the post scan to rows that
+# Soft tier: sketch / doodle / scribble (+ plurals), letter-boundary anchored so
+# "sketchbook" / "kadoodle" don't trip it. Deliberately conservative — recall is
+# secondary because the soft source doesn't train the head and the ring-loud audit
+# catches false positives.
+_SOFT_WIP_RE = re.compile(
+    r"(?<![A-Za-z])(?:sketch|sketches|doodle|doodles|scribble|scribbles)(?![A-Za-z])",
+    re.IGNORECASE,
+)
+
+# Coarse SQL prefilters for the backfill sweep — narrow the post scan to rows that
 # COULD match before the precise regex confirms. Case-insensitive ILIKE patterns.
-# MUST stay a SUPERSET of _WIP_RE (every regex match contains "wip" or
-# "work…progress") or the sweep would silently miss posts.
+# Each MUST stay a SUPERSET of its regex or the sweep would silently miss posts.
 WIP_TITLE_SQL_PREFILTER = ("%wip%", "%work%progress%")
+SOFT_WIP_TITLE_SQL_PREFILTER = ("%sketch%", "%doodle%", "%scribble%")
 
 # Chunk bulk inserts so a large sweep can't blow past psycopg's 65535-parameter
 # ceiling (3 params/row → ~21k rows max; 5k stays comfortably under).
@@ -51,12 +66,19 @@ _INSERT_CHUNK = 5000
 
 
 def matches_wip_title(title: str | None) -> bool:
-    """True when a post title explicitly marks it work-in-progress."""
+    """True when a post title explicitly marks it work-in-progress (HARD tier)."""
     if not title:
         return False
     return _WIP_RE.search(title) is not None
 
 
+def matches_soft_wip_title(title: str | None) -> bool:
+    """True when a title carries a SOFT WIP cue (sketch/doodle/scribble, #1474)."""
+    if not title:
+        return False
+    return _SOFT_WIP_RE.search(title) is not None
+
+
 def resolve_wip_tag_id(session: Session) -> int | None:
     """The seeded ``wip`` system tag's id (migration 0075), or None if absent."""
     return session.execute(
@@ -64,8 +86,10 @@ def resolve_wip_tag_id(session: Session) -> int | None:
     ).scalar_one_or_none()
 
 
-def apply_wip_image_tags(session: Session, image_ids, tag_id: int) -> int:
-    """Attach ``tag_id`` (source='wip_title') to each image id, idempotently —
+def apply_wip_image_tags(
+    session: Session, image_ids, tag_id: int, *, source: str = WIP_TITLE_SOURCE
+) -> int:
+    """Attach ``tag_id`` (stamped with ``source``) to each image id, idempotently —
     never disturbs an existing tag or its source. Returns the number of image_tag
     rows newly inserted. Does NOT commit.
 
@@ -92,7 +116,7 @@ def apply_wip_image_tags(session: Session, image_ids, tag_id: int) -> int:
         session.execute(
             pg_insert(image_tag)
             .values([
-                {"image_record_id": iid, "tag_id": tag_id, "source": WIP_TITLE_SOURCE}
+                {"image_record_id": iid, "tag_id": tag_id, "source": source}
                 for iid in to_insert
             ])
             .on_conflict_do_nothing(index_elements=["image_record_id", "tag_id"])

backend/app/tasks/maintenance.py

@@ -1047,6 +1047,41 @@ def cleanup_old_download_events() -> int:
         return result.rowcount or 0
 
 
+def _backfill_wip_tier(session, tag_id, prefilter, matcher, source) -> int:
+    """One keyset-paginated pass over posts whose title matches a WIP tier, applying
+    `tag_id` (stamped `source`) to their images. Shared by the hard + soft tiers
+    (#1458 / #1474). Coarse `prefilter` (ILIKE superset) narrows the scan; the precise
+    `matcher` confirms. Idempotent-additive (ON CONFLICT DO NOTHING). Returns the row
+    count newly applied."""
+    from ..models import Post
+    from ..models.image_provenance import ImageProvenance
+    from ..services.wip_title import apply_wip_image_tags
+
+    applied = 0
+    last_id = 0
+    while True:
+        rows = session.execute(
+            select(Post.id, Post.post_title)
+            .where(Post.id > last_id)
+            .where(Post.post_title.is_not(None))
+            .where(or_(*[Post.post_title.ilike(p) for p in prefilter]))
+            .order_by(Post.id.asc())
+            .limit(WIP_BACKFILL_PAGE)
+        ).all()
+        if not rows:
+            break
+        last_id = rows[-1][0]
+        match_ids = [pid for pid, title in rows if matcher(title)]
+        if match_ids:
+            image_ids = session.execute(
+                select(ImageProvenance.image_record_id)
+                .where(ImageProvenance.post_id.in_(match_ids))
+            ).scalars().all()
+            applied += apply_wip_image_tags(session, image_ids, tag_id, source=source)
+            session.commit()
+    return applied
+
+
 @celery.task(
     name="backend.app.tasks.maintenance.backfill_wip_title_tags",
     # Coarse-prefiltered scan over posts; the candidate set is small on a typical
@@ -1054,32 +1089,32 @@ def cleanup_old_download_events() -> int:
     soft_time_limit=1800, time_limit=2100,
 )
 def backfill_wip_title_tags() -> int:
-    """Scan EXISTING posts for explicit WIP titles and apply the `wip` system tag
-    to their images — the operator-triggered back-catalogue catch-up for
-    title-based WIP tagging (task #1458). New imports are tagged live by the
-    importer; this covers everything already in the library.
+    """Scan EXISTING posts for WIP titles and apply the `wip` system tag to their
+    images — the operator-triggered back-catalogue catch-up (task #1458 hard tier +
+    #1474 soft tier). New imports are tagged live by the importer; this covers the
+    existing library.
 
-    Keyset-paginated over posts (restart-safe). A coarse SQL prefilter narrows to
-    titles that COULD match; the precise regex (matches_wip_title) confirms.
-    Idempotent-additive (ON CONFLICT DO NOTHING) — never disturbs an existing tag.
+    HARD tier ("WIP"/"work in progress") always runs (the operator triggered the
+    scan); the SOFT tier (sketch/doodle, provisional source) runs only when
+    wip_soft_title_tagging_enabled, AFTER hard so a title matching both keeps the
+    trained hard tag (ON CONFLICT DO NOTHING). Keyset-paginated, restart-safe.
 
-    Deliberately NOT scheduled as a beat: a periodic re-run would re-apply to
-    matching posts and silently undo a manual WIP removal, so it stays an explicit
-    operator action (Settings → "Scan existing posts for WIP titles"). Returns the
-    number of image-tag rows newly applied.
+    Deliberately NOT scheduled as a beat: a periodic re-run would re-apply to matching
+    posts and silently undo a manual WIP removal, so it stays an explicit operator
+    action (Settings → "Scan existing posts for WIP titles"). Returns rows applied.
     """
-    from ..models import Post
-    from ..models.image_provenance import ImageProvenance
+    from ..models import ImportSettings
     from ..services.wip_title import (
+        SOFT_WIP_TITLE_SQL_PREFILTER,
+        WIP_TITLE_SOFT_SOURCE,
+        WIP_TITLE_SOURCE,
         WIP_TITLE_SQL_PREFILTER,
-        apply_wip_image_tags,
+        matches_soft_wip_title,
         matches_wip_title,
         resolve_wip_tag_id,
     )
 
     SessionLocal = _sync_session_factory()
-    applied = 0
-    last_id = 0
     with SessionLocal() as session:
         tag_id = resolve_wip_tag_id(session)
         if tag_id is None:
@@ -1087,30 +1122,16 @@ def backfill_wip_title_tags() -> int:
                 "backfill_wip_title_tags: no `wip` system tag present; nothing to do"
             )
             return 0
-        like_a, like_b = WIP_TITLE_SQL_PREFILTER
-        while True:
-            rows = session.execute(
-                select(Post.id, Post.post_title)
-                .where(Post.id > last_id)
-                .where(Post.post_title.is_not(None))
-                .where(or_(
-                    Post.post_title.ilike(like_a),
-                    Post.post_title.ilike(like_b),
-                ))
-                .order_by(Post.id.asc())
-                .limit(WIP_BACKFILL_PAGE)
-            ).all()
-            if not rows:
-                break
-            last_id = rows[-1][0]
-            match_ids = [pid for pid, title in rows if matches_wip_title(title)]
-            if match_ids:
-                image_ids = session.execute(
-                    select(ImageProvenance.image_record_id)
-                    .where(ImageProvenance.post_id.in_(match_ids))
-                ).scalars().all()
-                applied += apply_wip_image_tags(session, image_ids, tag_id)
-                session.commit()
+        settings = ImportSettings.load_sync(session)
+        applied = _backfill_wip_tier(
+            session, tag_id, WIP_TITLE_SQL_PREFILTER, matches_wip_title,
+            WIP_TITLE_SOURCE,
+        )
+        if settings.wip_soft_title_tagging_enabled:
+            applied += _backfill_wip_tier(
+                session, tag_id, SOFT_WIP_TITLE_SQL_PREFILTER, matches_soft_wip_title,
+                WIP_TITLE_SOFT_SOURCE,
+            )
     if applied:
         log.info("backfill_wip_title_tags: applied wip to %d image(s)", applied)
     return applied

backend/app/tasks/ml.py

@@ -599,18 +599,54 @@ def scheduled_ccip_auto_apply() -> str:
     soft_time_limit=1800, time_limit=2100,
 )
 def scheduled_presentation_auto_apply() -> str:
-    """Auto-hide presentation chrome (banner / editor screenshot) on a daily
-    passive sweep (#141). No-op unless presentation_auto_apply_enabled. Idempotent
-    — already-hidden images are skipped — so an interrupted run simply re-runs next
-    cycle (that IS the recovery). Wall-clock bounded by the task time limits."""
-    from ..services.ml.heads import presentation_auto_apply_sweep
+    """Auto-hide presentation chrome (banner) on a daily passive sweep (#141).
+    No-op unless presentation_auto_apply_enabled. Idempotent — already-tagged images
+    are skipped — so an interrupted run simply re-runs next cycle (that IS the
+    recovery). Wall-clock bounded by the task time limits."""
+    from ..services.ml.heads import system_tag_auto_apply_sweep
 
     SessionLocal = _sync_session_factory()
     with SessionLocal() as session:
-        result = presentation_auto_apply_sweep(session)
+        result = system_tag_auto_apply_sweep(session, mode="chrome")
     return f"applied={result['n_applied']} flagged={result['n_flagged']}"
 
 
+@celery.task(
+    name="backend.app.tasks.ml.scheduled_process_auto_apply",
+    soft_time_limit=1800, time_limit=2100,
+)
+def scheduled_process_auto_apply() -> str:
+    """Auto-apply the PROCESS system tags (wip / editor screenshot) on a daily
+    passive sweep (#1464) — provisional source, ring-loud review guard, image stays
+    VISIBLE. No-op unless process_auto_apply_enabled (opt-in). Idempotent —
+    already-tagged/rejected images are skipped — so an interrupted run just re-runs
+    next cycle (the recovery). Wall-clock bounded by the task time limits."""
+    from ..services.ml.heads import system_tag_auto_apply_sweep
+
+    SessionLocal = _sync_session_factory()
+    with SessionLocal() as session:
+        result = system_tag_auto_apply_sweep(session, mode="process")
+    return f"applied={result['n_applied']} flagged={result['n_flagged']}"
+
+
+@celery.task(
+    name="backend.app.tasks.ml.scheduled_soft_wip_conflict_audit",
+    soft_time_limit=1800, time_limit=2100,
+)
+def scheduled_soft_wip_conflict_audit() -> str:
+    """Ring-loud audit over the SOFT WIP-title cohort (#1474) — flag sketch/doodle
+    auto-tags that ALSO look like real content for review. No-op when there are no
+    content heads; idempotent (already-flagged images skipped). Runs regardless of
+    the process-sweep toggle, since soft-title tags come from the importer, not that
+    sweep. Wall-clock bounded by the task time limits."""
+    from ..services.ml.heads import soft_wip_conflict_audit
+
+    SessionLocal = _sync_session_factory()
+    with SessionLocal() as session:
+        result = soft_wip_conflict_audit(session)
+    return f"scanned={result['n_scanned']} flagged={result['n_flagged']}"
+
+
 @celery.task(name="backend.app.tasks.ml.prune_presentation_reviews")
 def prune_presentation_reviews() -> str:
     """Retention (rule 89): drop RESOLVED presentation-review flags older than 30

frontend/src/components/gallery/HiddenReviewStrip.vue

@@ -1,14 +1,15 @@
 <template>
-  <!-- Auto-hidden chrome that ALSO looked like real content — surfaced PROACTIVELY
-       atop the gallery whenever there's something to review (NOT gated on the
-       Show-hidden toggle, so misfires can't go unnoticed), most-concerning first,
-       with keep / un-hide (#141). Renders nothing when there's nothing to review. -->
-  <section v-if="items.length" class="fc-review" aria-label="Hidden images to review">
+  <!-- System-tag auto-applies (chrome hides / process WIP tags) that ALSO looked
+       like real content — surfaced PROACTIVELY atop the gallery whenever there's
+       something to review (NOT gated on the Show-hidden toggle, so misfires can't
+       go unnoticed), most-concerning first, with keep / remove (#141, #1464).
+       Renders nothing when there's nothing to review. -->
+  <section v-if="items.length" class="fc-review" aria-label="Auto-tagged images to review">
     <div class="fc-review__head">
       <v-icon size="18" color="warning">mdi-alert-outline</v-icon>
       <span class="fc-review__title">
-        {{ items.length }} auto-hidden {{ items.length === 1 ? 'image' : 'images' }}
-        may be real content — review before they stay hidden
+        {{ items.length }} auto-tagged {{ items.length === 1 ? 'image' : 'images' }}
+        may be real content — review
       </span>
     </div>
     <div class="fc-review__cards">
@@ -26,16 +27,16 @@
           >
             also looks like <strong>{{ it.conflict_name || 'content' }}</strong>
           </div>
-          <div class="fc-review-card__tag">hidden as {{ it.tag_name }}</div>
+          <div class="fc-review-card__tag">{{ tagLine(it) }}</div>
           <div class="fc-review-card__acts">
             <button
               type="button" class="fc-review-btn fc-review-btn--keep"
               :disabled="busy.includes(keyOf(it))" @click="resolve(it, 'keep')"
-            >Keep hidden</button>
+            >{{ keepLabel(it) }}</button>
             <button
               type="button" class="fc-review-btn fc-review-btn--unhide"
               :disabled="busy.includes(keyOf(it))" @click="resolve(it, 'unhide')"
-            >Un-hide</button>
+            >{{ removeLabel(it) }}</button>
           </div>
         </div>
       </div>
@@ -54,6 +55,11 @@ const items = ref([])
 const busy = ref([])
 
 function keyOf(it) { return `${it.image_id}:${it.tag_id}` }
+// Chrome flags hide the image (keep-hidden / un-hide); process flags leave it
+// visible and just tagged (keep-tag / remove-tag). Same endpoints, different words.
+function tagLine(it) { return (it.mode === 'process' ? 'auto-tagged ' : 'hidden as ') + it.tag_name }
+function keepLabel(it) { return it.mode === 'process' ? 'Keep tag' : 'Keep hidden' }
+function removeLabel(it) { return it.mode === 'process' ? 'Remove tag' : 'Un-hide' }
 
 async function load() {
   // Fetched unconditionally on mount — the strip prompts for pending misfires
@@ -71,7 +77,8 @@ async function resolve(it, action) {
     await api.post(`/api/gallery/hidden-review/${it.image_id}/${it.tag_id}/${action}`)
     items.value = items.value.filter((x) => keyOf(x) !== k)
     if (action === 'unhide') {
-      toast({ text: `Un-hidden — “${it.tag_name}” removed; it'll train the head`, type: 'success' })
+      const verb = it.mode === 'process' ? 'Removed' : 'Un-hidden'
+      toast({ text: `${verb} — “${it.tag_name}” removed; it'll train the head`, type: 'success' })
     }
   } catch (e) {
     toast({

frontend/src/components/settings/HeadsCard.vue

@@ -171,13 +171,13 @@
         />
       </div>
       <p class="fc-muted text-body-2 mb-3">
-        Auto-hide banners and editor screenshots from the gallery once a head has
-        learned them (≥ {{ minPositives }} examples) and clears
+        Auto-hide <code>banner</code> chrome from the gallery once a head has
+        learned it (≥ {{ minPositives }} examples) and clears
         {{ Math.round((presentationThresholdInput || 0) * 100) }}% confidence.
-        <code>wip</code> is never auto-hidden. If a hidden image also looks like
-        real content (≥ {{ Math.round((presentationConflictInput || 0) * 100) }}%
-        on a content tag), it's flagged in the Hidden view instead of buried.
-        Every auto-hide is reversible.
+        (<code>wip</code> and <code>editor screenshot</code> are handled by the
+        process auto-tagger below.) If a hidden image also looks like real content
+        (≥ {{ Math.round((presentationConflictInput || 0) * 100) }}% on a content
+        tag), it's flagged for review instead of buried. Every auto-hide is reversible.
       </p>
       <div class="d-flex mb-3" style="gap: 12px;">
         <v-text-field
@@ -195,6 +195,43 @@
       </div>
     </div>
 
+    <!-- Process auto-tagging (#1464): wip / editor screenshot -->
+    <div class="fc-auto mt-6">
+      <div class="d-flex align-center mb-1" style="gap: 10px;">
+        <v-icon size="18" color="accent">mdi-progress-wrench</v-icon>
+        <span class="fc-section-h">Auto-tag work-in-progress</span>
+        <v-switch
+          v-model="processEnabled" :loading="settingBusy" hide-details
+          density="compact" color="success" class="ml-auto"
+          @update:model-value="onToggleProcess"
+        />
+      </div>
+      <p class="fc-muted text-body-2 mb-3">
+        Auto-tag <code>wip</code> and <code>editor screenshot</code> process art
+        once a head has learned them (≥ {{ minPositives }} examples) and clears
+        {{ Math.round((processThresholdInput || 0) * 100) }}% confidence. These stay
+        <strong>visible</strong> in the gallery — the tag just keeps them out of
+        training and the Explore rabbit-hole. Off by default. If a tagged image also
+        looks like real content (≥ {{ Math.round((processConflictInput || 0) * 100) }}%
+        on a content tag), it's flagged for review. Learns only from your titles +
+        manual tags, never its own guesses — so it can't run away. Every tag reversible.
+      </p>
+      <div class="d-flex mb-3" style="gap: 12px;">
+        <v-text-field
+          v-model.number="processThresholdInput" label="Tag confidence"
+          type="number" min="0.5" max="0.999" step="0.01" density="compact"
+          hide-details style="max-width: 200px;" :disabled="settingBusy"
+          @change="onSaveProcess"
+        />
+        <v-text-field
+          v-model.number="processConflictInput" label="Flag if content ≥"
+          type="number" min="0" max="1" step="0.05" density="compact"
+          hide-details style="max-width: 200px;" :disabled="settingBusy"
+          @change="onSaveProcess"
+        />
+      </div>
+    </div>
+
     <!-- Performance / tuning -->
     <div v-if="metricsConcepts.length" class="mt-5">
       <div class="fc-section-h mb-1">How auto-apply is landing</div>
@@ -258,6 +295,9 @@ let autoTimer = null
 const presentationEnabled = ref(true)
 const presentationThresholdInput = ref(0.90)
 const presentationConflictInput = ref(0.50)
+const processEnabled = ref(false)
+const processThresholdInput = ref(0.90)
+const processConflictInput = ref(0.50)
 
 const autoRunning = computed(() => autoStatus.value?.running_id != null)
 const lastSweep = computed(() =>
@@ -292,6 +332,9 @@ onMounted(async () => {
     presentationEnabled.value = s.presentation_auto_apply_enabled ?? true
     presentationThresholdInput.value = s.presentation_auto_apply_threshold ?? 0.90
     presentationConflictInput.value = s.presentation_conflict_threshold ?? 0.50
+    processEnabled.value = s.process_auto_apply_enabled ?? false
+    processThresholdInput.value = s.process_auto_apply_threshold ?? 0.90
+    processConflictInput.value = s.process_conflict_threshold ?? 0.50
   } catch { /* non-fatal */ }
   await refresh()
   if (running.value) startPoll()
@@ -402,6 +445,32 @@ async function onSavePresentation() {
     settingBusy.value = false
   }
 }
+
+async function onToggleProcess(val) {
+  settingBusy.value = true
+  try {
+    await mlSettings.patchSettings({ process_auto_apply_enabled: !!val })
+    toast({ text: val ? 'WIP auto-tag on' : 'WIP auto-tag off', type: 'success' })
+  } catch (e) {
+    processEnabled.value = !val   // revert the switch
+    toast({ text: `Could not update: ${e.message}`, type: 'error' })
+  } finally {
+    settingBusy.value = false
+  }
+}
+async function onSaveProcess() {
+  settingBusy.value = true
+  try {
+    await mlSettings.patchSettings({
+      process_auto_apply_threshold: Number(processThresholdInput.value),
+      process_conflict_threshold: Number(processConflictInput.value),
+    })
+  } catch (e) {
+    toast({ text: `Could not save: ${e.message}`, type: 'error' })
+  } finally {
+    settingBusy.value = false
+  }
+}
 function onPreview() { startSweep(true) }
 function onApplyNow() { startSweep(false) }
 async function startSweep(dryRun) {

frontend/src/components/settings/ImportFiltersForm.vue

@@ -99,6 +99,17 @@
           the Explore browse. Applies to new imports; run the scan below to catch
           posts already in your library.
         </div>
+        <v-switch
+          v-model="local.wip_soft_title_tagging_enabled"
+          label="Also tag “sketch” / “doodle” titles (lower precision)"
+          density="compact" hide-details color="primary" @change="save"
+        />
+        <div class="fc-help mb-3">
+          Extends the above to softer cues (<code>sketch</code>, <code>doodle</code>,
+          <code>scribble</code>). These stay <strong>visible</strong> and never train
+          the tagging model — a daily audit flags any that actually look like finished
+          art for review. Off by default.
+        </div>
         <v-btn
           variant="tonal" color="primary" size="small"
           :loading="store.wipScanBusy" prepend-icon="mdi-magnify"
@@ -139,6 +150,7 @@ const local = reactive({
   skip_single_color: false, single_color_threshold: 0.95,
   phash_threshold: 10,
   wip_title_tagging_enabled: true,
+  wip_soft_title_tagging_enabled: false,
 })
 
 watch(() => store.settings, (s) => { if (s) Object.assign(local, s) }, { immediate: true })

frontend/src/stores/explore.js

@@ -25,6 +25,10 @@ export const useExploreStore = defineStore('explore', () => {
   const cursor = ref(-1)
   const loading = ref(false)
   const error = ref(null)
+  // Reach (#1476): how far the walk reaches past the anchor's immediate cluster.
+  // 0 = nearest (can get stuck in a dense signature); ~0.4 default mixes in
+  // mid-far escape routes so the walk diversifies without hitting "Random image".
+  const reach = ref(0.4)
 
   const inflight = useInflightToken()
 
@@ -46,7 +50,13 @@ export const useExploreStore = defineStore('explore', () => {
         const body = await api.get('/api/gallery/similar', {
           // exclude_wip: keep work-in-progress out of the Explore rabbit-hole
           // (the gallery's own "similar" button still shows it) — operator 2026-07-08.
-          params: { similar_to: numId, limit: NEIGHBOR_LIMIT, exclude_wip: 1 },
+          // reach + exclude_ids (#1476): reach past the dense cluster + never re-serve
+          // an already-walked image, so the walk keeps moving instead of getting stuck.
+          params: {
+            similar_to: numId, limit: NEIGHBOR_LIMIT, exclude_wip: 1,
+            reach: reach.value,
+            exclude_ids: breadcrumb.value.map((c) => c.id).join(','),
+          },
         })
         if (!t.isCurrent()) return
         neighbors.value = body.images || []
@@ -113,6 +123,14 @@ export const useExploreStore = defineStore('explore', () => {
     loading.value = false
   }
 
+  // Change how far the walk reaches and re-fetch the current anchor's neighbours
+  // with the new setting (the anchor + trail are unchanged — only the grid varies).
+  function setReach (v) {
+    reach.value = Math.max(0, Math.min(1, Number(v)))
+    const id = anchor.value?.id
+    if (id != null) anchorOn(id)
+  }
+
   // --- TagPanel "host" surface ---------------------------------------------
   // The anchor IS the current image (same /api/gallery/image/<id> payload the
   // modal uses), so these mirror the modal store's tag-CRUD, targeting the
@@ -189,8 +207,8 @@ export const useExploreStore = defineStore('explore', () => {
   function close () {}
 
   return {
-    anchor, neighbors, breadcrumb, cursor, loading, error, NEIGHBOR_LIMIT,
-    anchorOn, reset, backTarget, forwardTarget,
+    anchor, neighbors, breadcrumb, cursor, loading, error, NEIGHBOR_LIMIT, reach,
+    anchorOn, reset, backTarget, forwardTarget, setReach,
     // host surface
     current, currentImageId,
     reloadTags, addExistingTag, removeTag, createAndAdd, close,

frontend/src/views/ExploreView.vue

@@ -31,6 +31,20 @@
           <img :src="c.thumbnail_url" alt="" loading="lazy" />
         </button>
         <div class="fc-ex__trail-actions">
+          <!-- Reach (#1476): how far each step reaches past the anchor's cluster.
+               Raise it to break out of a dense signature without hitting Random. -->
+          <div
+            class="fc-ex__reach"
+            title="How far each step reaches — raise it to escape a dense cluster without going fully random"
+          >
+            <v-icon size="16" color="accent">mdi-map-marker-distance</v-icon>
+            <v-slider
+              :model-value="store.reach" @end="store.setReach"
+              :min="0" :max="1" :step="0.2" hide-details density="compact"
+              color="accent" class="fc-ex__reach-slider"
+            />
+            <span class="fc-muted fc-ex__reach-label">{{ reachLabel }}</span>
+          </div>
           <!-- Active retrain right where you tag: fold the +/- you just gave
                into the heads without a trip to Settings (the nightly beat is the
                passive cadence). -->
@@ -153,6 +167,12 @@ const modal = useModalStore()
 
 const anchorId = computed(() => route.params.imageId || null)
 const isVideo = computed(() => !!store.anchor?.mime?.startsWith('video/'))
+const reachLabel = computed(() => {
+  const r = store.reach
+  if (r <= 0.15) return 'Near'
+  if (r <= 0.55) return 'Varied'
+  return 'Far'
+})
 
 // #1206: hovering a suggestion in the rail highlights the crop it came from on
 // the anchor image (same provide/inject as the modal viewer).
@@ -296,6 +316,14 @@ onUnmounted(() => {
 .fc-ex__trail-actions {
   margin-left: auto; display: flex; align-items: center; gap: 4px; flex: 0 0 auto;
 }
+.fc-ex__reach {
+  display: flex; align-items: center; gap: 6px;
+  margin-right: 8px;
+}
+.fc-ex__reach-slider { width: 96px; }
+.fc-ex__reach-label {
+  font-size: 12px; min-width: 44px; text-align: left;
+}
 
 /* The three panes fill the remaining height; each scrolls on its own.
    grid-template-rows: minmax(0, 1fr) BOUNDS the single row to the container

tests/test_gallery_service.py

@@ -82,24 +82,28 @@ async def _system_tag(db, name):
 
 @pytest.mark.asyncio
 async def test_scroll_hides_presentation_chrome_by_default(db):
-    # banner / editor screenshot (presentation system tags) are hidden from the
-    # default gallery; wip (also a system tag) is NOT — it's real, in-progress
-    # art (milestone 141). Seeded system tags survive the harness TRUNCATE.
-    imgs = await _seed_images(db, 3, sha_prefix="p")
+    # banner (chrome) is hidden from the default gallery; wip AND editor screenshot
+    # (the PROCESS system tags) are NOT — they're real art / process shots that stay
+    # visible (milestone 141 + #1464). Seeded system tags survive the harness TRUNCATE.
+    imgs = await _seed_images(db, 4, sha_prefix="p")
     banner = await _system_tag(db, "banner")
     wip = await _system_tag(db, "wip")
+    editor = await _system_tag(db, "editor screenshot")
     await db.execute(image_tag.insert().values(
         image_record_id=imgs[0].id, tag_id=banner.id, source="manual"))
     await db.execute(image_tag.insert().values(
         image_record_id=imgs[1].id, tag_id=wip.id, source="manual"))
+    await db.execute(image_tag.insert().values(
+        image_record_id=imgs[3].id, tag_id=editor.id, source="manual"))
     await db.flush()
     svc = GalleryService(db)
 
-    # Default: the banner image is hidden; the wip image + the plain image stay.
+    # Default: only the banner image is hidden; wip + editor + plain all stay.
     default_ids = {i.id for i in (await svc.scroll(cursor=None, limit=10)).images}
     assert imgs[0].id not in default_ids      # banner hidden
     assert imgs[1].id in default_ids          # wip visible
     assert imgs[2].id in default_ids          # plain visible
+    assert imgs[3].id in default_ids          # editor screenshot visible (#1464)
 
     # include_hidden surfaces the banner image (the Hidden view).
     shown = {i.id for i in (

tests/test_gallery_similar.py

@@ -170,6 +170,23 @@ async def test_similar_respects_limit(db):
     assert len(res) == 2
 
 
+@pytest.mark.asyncio
+async def test_similar_exclude_ids_drops_walked(db):
+    """Explore passes its breadcrumb as exclude_ids so already-walked images
+    aren't re-served as neighbours (#1476). reach>0 runs cleanly too (small pool
+    → the sampler passes through)."""
+    src = await _img(db, 1, _vec(1, 0))
+    walked = await _img(db, 2, _vec(1, 0.05))
+    fresh = await _img(db, 3, _vec(1, 0.3))
+    svc = GalleryService(db)
+    res = await svc.similar(src.id, limit=10, exclude_ids=[walked.id])
+    ids = {i.id for i in res}
+    assert walked.id not in ids
+    assert fresh.id in ids
+    res_reach = await svc.similar(src.id, limit=10, reach=1.0)
+    assert fresh.id in {i.id for i in res_reach}
+
+
 # --- API ---
 
 @pytest.mark.asyncio

tests/test_presentation_auto_apply.py

@@ -15,7 +15,7 @@ from backend.app.models import (
 from backend.app.models.tag import image_tag
 from backend.app.services.ml.heads import (
     auto_apply_sweep,
-    presentation_auto_apply_sweep,
+    system_tag_auto_apply_sweep,
 )
 
 pytestmark = pytest.mark.integration
@@ -70,7 +70,7 @@ def test_presentation_sweep_hides_chrome(db_sync):
     _head(db_sync, banner.id, 0, weight=3.0)
     img = _img(db_sync, "a" * 64, _emb(0))
     db_sync.commit()
-    res = presentation_auto_apply_sweep(db_sync)
+    res = system_tag_auto_apply_sweep(db_sync, mode="chrome")
     assert res["n_applied"] == 1
     assert _source(db_sync, img.id, banner.id) == "presentation_auto"
 
@@ -86,7 +86,7 @@ def test_presentation_sweep_hard_skips_valued_image(db_sync):
     db_sync.execute(image_tag.insert().values(
         image_record_id=img.id, tag_id=content.id, source="manual"))
     db_sync.commit()
-    res = presentation_auto_apply_sweep(db_sync)
+    res = system_tag_auto_apply_sweep(db_sync, mode="chrome")
     assert res["n_applied"] == 0
     assert _source(db_sync, img.id, banner.id) is None
 
@@ -102,7 +102,7 @@ def test_presentation_sweep_flags_conflict(db_sync):
     _head(db_sync, content.id, 0, weight=1.0)     # content head also fires
     img = _img(db_sync, "c" * 64, _emb(0))
     db_sync.commit()
-    res = presentation_auto_apply_sweep(db_sync)
+    res = system_tag_auto_apply_sweep(db_sync, mode="chrome")
     assert res["n_applied"] == 1
     assert res["n_flagged"] == 1
     assert _source(db_sync, img.id, banner.id) == "presentation_auto"
@@ -123,7 +123,7 @@ def test_presentation_sweep_disabled_is_noop(db_sync):
     _head(db_sync, banner.id, 0, weight=3.0)
     img = _img(db_sync, "d" * 64, _emb(0))
     db_sync.commit()
-    res = presentation_auto_apply_sweep(db_sync)
+    res = system_tag_auto_apply_sweep(db_sync, mode="chrome")
     assert res["n_applied"] == 0
     assert _source(db_sync, img.id, banner.id) is None
 
@@ -134,7 +134,7 @@ def test_presentation_sweep_ignores_wip(db_sync):
     _head(db_sync, wip.id, 0, weight=3.0)
     img = _img(db_sync, "e" * 64, _emb(0))
     db_sync.commit()
-    res = presentation_auto_apply_sweep(db_sync)
+    res = system_tag_auto_apply_sweep(db_sync, mode="chrome")
     assert res["n_applied"] == 0
     assert _source(db_sync, img.id, wip.id) is None
 

tests/test_process_auto_apply.py

@@ -0,0 +1,188 @@
+"""Process-group auto-apply sweep (#1464): wip / editor screenshot auto-tag at a
+flat threshold with a PROVISIONAL source (`process_auto`) so the head never trains
+on its own output, and stay VISIBLE (unlike chrome). Mirrors the chrome guards.
+numpy-only (no sklearn), tested directly via the sync session."""
+import pytest
+from sqlalchemy import select
+
+from backend.app.models import (
+    ImageRecord,
+    MLSettings,
+    PresentationReview,
+    Tag,
+    TagHead,
+    TagKind,
+)
+from backend.app.models.tag import image_tag
+from backend.app.services.ml.heads import system_tag_auto_apply_sweep
+from backend.app.services.ml.training_data import _ids_with_tag
+
+pytestmark = pytest.mark.integration
+
+
+def _emb(slot: int) -> list[float]:
+    v = [0.0] * 1152
+    v[slot] = 3.0
+    return v
+
+
+def _img(db, sha: str, emb) -> ImageRecord:
+    img = ImageRecord(
+        path=f"/images/{sha}.jpg", sha256=sha, size_bytes=1, mime="image/jpeg",
+        width=1, height=1, origin="imported_filesystem",
+        integrity_status="unknown", siglip_embedding=emb,
+    )
+    db.add(img)
+    db.flush()
+    return img
+
+
+def _head(db, tag_id: int, slot: int, *, weight=1.0):
+    s = db.execute(select(MLSettings).where(MLSettings.id == 1)).scalar_one()
+    w = [0.0] * 1152
+    w[slot] = weight
+    db.add(TagHead(
+        tag_id=tag_id, embedding_version=s.embedder_model_version,
+        weights=w, bias=0.0, suggest_threshold=0.5, auto_apply_threshold=0.5,
+        n_pos=60, n_neg=90, ap=0.9, precision_cv=0.98, recall=0.7,
+    ))
+
+
+def _system_tag(db, name):
+    return db.execute(
+        select(Tag).where(Tag.is_system.is_(True), Tag.name == name)
+    ).scalar_one()
+
+
+def _enable_process(db):
+    # process auto-apply is opt-in (default False) — turn it on for these tests.
+    db.execute(select(MLSettings).where(MLSettings.id == 1)).scalar_one().process_auto_apply_enabled = True
+
+
+def _source(db, image_id, tag_id):
+    return db.execute(
+        select(image_tag.c.source)
+        .where(image_tag.c.image_record_id == image_id)
+        .where(image_tag.c.tag_id == tag_id)
+    ).scalar_one_or_none()
+
+
+def test_process_sweep_applies_wip_and_editor(db_sync):
+    _enable_process(db_sync)
+    wip = _system_tag(db_sync, "wip")
+    editor = _system_tag(db_sync, "editor screenshot")
+    _head(db_sync, wip.id, 0, weight=3.0)
+    _head(db_sync, editor.id, 1, weight=3.0)
+    w_img = _img(db_sync, "a" * 64, _emb(0))
+    e_img = _img(db_sync, "b" * 64, _emb(1))
+    db_sync.commit()
+    res = system_tag_auto_apply_sweep(db_sync, mode="process")
+    assert res["n_applied"] == 2
+    assert _source(db_sync, w_img.id, wip.id) == "process_auto"
+    assert _source(db_sync, e_img.id, editor.id) == "process_auto"
+
+
+def test_process_sweep_disabled_by_default_is_noop(db_sync):
+    # process_auto_apply_enabled defaults False (opt-in) — no enable = no-op.
+    wip = _system_tag(db_sync, "wip")
+    _head(db_sync, wip.id, 0, weight=3.0)
+    img = _img(db_sync, "c" * 64, _emb(0))
+    db_sync.commit()
+    res = system_tag_auto_apply_sweep(db_sync, mode="process")
+    assert res["n_applied"] == 0
+    assert _source(db_sync, img.id, wip.id) is None
+
+
+def test_process_sweep_skips_valued_image(db_sync):
+    # Guard 1: never auto-apply to an image the operator already content-tagged.
+    _enable_process(db_sync)
+    wip = _system_tag(db_sync, "wip")
+    _head(db_sync, wip.id, 0, weight=3.0)
+    content = Tag(name="mychar", kind=TagKind.character)
+    db_sync.add(content)
+    db_sync.flush()
+    img = _img(db_sync, "d" * 64, _emb(0))
+    db_sync.execute(image_tag.insert().values(
+        image_record_id=img.id, tag_id=content.id, source="manual"))
+    db_sync.commit()
+    res = system_tag_auto_apply_sweep(db_sync, mode="process")
+    assert res["n_applied"] == 0
+    assert _source(db_sync, img.id, wip.id) is None
+
+
+def test_process_sweep_flags_conflict_with_process_mode(db_sync):
+    # Guard 2: also scores high on a content head → still applied, but flagged
+    # for review with mode='process' (the ring-loud guard).
+    _enable_process(db_sync)
+    wip = _system_tag(db_sync, "wip")
+    _head(db_sync, wip.id, 0, weight=3.0)
+    content = Tag(name="looksreal", kind=TagKind.general)
+    db_sync.add(content)
+    db_sync.flush()
+    _head(db_sync, content.id, 0, weight=1.0)
+    img = _img(db_sync, "e" * 64, _emb(0))
+    db_sync.commit()
+    res = system_tag_auto_apply_sweep(db_sync, mode="process")
+    assert res["n_applied"] == 1
+    assert res["n_flagged"] == 1
+    flag = db_sync.execute(
+        select(PresentationReview).where(
+            PresentationReview.image_record_id == img.id,
+            PresentationReview.tag_id == wip.id,
+        )
+    ).scalar_one()
+    assert flag.mode == "process"
+    assert flag.conflict_tag_id == content.id
+
+
+def test_process_auto_source_never_trains_head(db_sync):
+    # The runaway break: provisional wip tags (process sweep 'process_auto', soft
+    # title 'wip_title_soft') are NOT training positives; a HARD title-heuristic /
+    # manual one IS. So the head learns only from trusted labels, never its own
+    # output or the low-precision sketch/doodle tier (#1464 + #1474).
+    wip = _system_tag(db_sync, "wip")
+    auto_img = _img(db_sync, "f" * 64, _emb(0))
+    soft_img = _img(db_sync, "9" * 64, _emb(2))
+    title_img = _img(db_sync, "0" * 64, _emb(1))
+    db_sync.execute(image_tag.insert().values(
+        image_record_id=auto_img.id, tag_id=wip.id, source="process_auto"))
+    db_sync.execute(image_tag.insert().values(
+        image_record_id=soft_img.id, tag_id=wip.id, source="wip_title_soft"))
+    db_sync.execute(image_tag.insert().values(
+        image_record_id=title_img.id, tag_id=wip.id, source="wip_title"))
+    db_sync.commit()
+    positives = set(_ids_with_tag(db_sync, wip.id))
+    assert title_img.id in positives       # trusted HARD label trains the head
+    assert auto_img.id not in positives    # its own auto-applied output does NOT
+    assert soft_img.id not in positives    # low-precision soft tier does NOT
+
+
+def test_soft_wip_conflict_audit_flags_ring_loud(db_sync):
+    # A soft-tagged image (sketch/doodle title) that ALSO scores high on a content
+    # head is probably finished art mis-tagged — flagged for review; a quiet one is not.
+    from backend.app.services.ml.heads import soft_wip_conflict_audit
+
+    s = db_sync.execute(select(MLSettings).where(MLSettings.id == 1)).scalar_one()
+    s.process_conflict_threshold = 0.6
+    wip = _system_tag(db_sync, "wip")
+    content = Tag(name="looksreal", kind=TagKind.general)
+    db_sync.add(content)
+    db_sync.flush()
+    _head(db_sync, content.id, 0, weight=1.0)   # sigmoid(1)=0.73 > 0.6 conflict
+    ring = _img(db_sync, "1" * 64, _emb(0))      # scores on the content head
+    quiet = _img(db_sync, "2" * 64, _emb(5))     # orthogonal → 0.5 < 0.6
+    for img in (ring, quiet):
+        db_sync.execute(image_tag.insert().values(
+            image_record_id=img.id, tag_id=wip.id, source="wip_title_soft"))
+    db_sync.commit()
+
+    res = soft_wip_conflict_audit(db_sync)
+    assert res["n_flagged"] == 1
+    flag = db_sync.execute(
+        select(PresentationReview).where(PresentationReview.image_record_id == ring.id)
+    ).scalar_one()
+    assert flag.mode == "process"
+    assert flag.conflict_tag_id == content.id
+    assert db_sync.execute(
+        select(PresentationReview).where(PresentationReview.image_record_id == quiet.id)
+    ).scalar_one_or_none() is None

tests/test_reach_sample.py

@@ -0,0 +1,32 @@
+"""Explore reach sampler (#1476) — pure unit tests for `_reach_sample`, which picks
+a distance-rank spread so the neighbour pool handed to MMR spans near→mid-far and
+the walk can escape a dense cluster. No DB. `_reach_sample` only indexes rows, so
+plain ints stand in for the (ImageRecord, ...) tuples."""
+from backend.app.services.gallery_service import _reach_sample
+
+
+def test_reach_zero_or_negative_passes_through():
+    rows = list(range(1000))
+    assert _reach_sample(rows, 40, 0.0) is rows
+    assert _reach_sample(rows, 40, -1.0) is rows
+
+
+def test_small_pool_passes_through():
+    # n <= want (limit*8 = 320) → nothing to reach into.
+    rows = list(range(50))
+    assert _reach_sample(rows, 40, 1.0) is rows
+
+
+def test_higher_reach_reaches_deeper_ranks():
+    rows = list(range(1000))
+    near = _reach_sample(rows, 40, 0.2)
+    far = _reach_sample(rows, 40, 1.0)
+    # Both keep the nearest rank (stride starts at 0) so you can still tag the cluster.
+    assert near[0] == 0
+    assert far[0] == 0
+    # But higher reach samples genuinely farther ranks.
+    assert max(far) > max(near)
+    assert max(far) >= 900          # reach=1 spans (almost) the whole pool
+    assert max(near) <= 550         # reach=0.2 stays in the near half
+    # Never runs past the pool.
+    assert max(far) <= len(rows) - 1

tests/test_wip_title.py

@@ -8,7 +8,7 @@ negative cases (substrings like ``swipe`` / ``wiped``) are the load-bearing ones
 """
 import pytest
 
-from backend.app.services.wip_title import matches_wip_title
+from backend.app.services.wip_title import matches_soft_wip_title, matches_wip_title
 
 
 @pytest.mark.parametrize("title", [
@@ -44,6 +44,33 @@ def test_matches_positive(title):
     "finished at last",
     "Kawips diner",                 # 'wip' mid-word
     "swipright",
+    "quick sketch of Nami",         # soft cue — NOT a HARD WIP match
 ])
 def test_matches_negative(title):
     assert matches_wip_title(title) is False
+
+
+@pytest.mark.parametrize("title", [
+    "quick sketch",
+    "Nami sketch",
+    "morning doodle",
+    "some doodles",
+    "sketches from today",
+    "a little scribble",
+    "SKETCH",
+])
+def test_soft_matches_positive(title):
+    assert matches_soft_wip_title(title) is True
+
+
+@pytest.mark.parametrize("title", [
+    None,
+    "",
+    "sketchbook tour",              # 'sketch' inside sketchbook — must NOT match
+    "kadoodle mascot",              # 'doodle' mid-word
+    "the final piece",
+    "WIP",                          # a HARD cue is not a SOFT cue
+    "prescribed colours",           # 'scrib' inside prescribed — must NOT match
+])
+def test_soft_matches_negative(title):
+    assert matches_soft_wip_title(title) is False

tests/test_wip_title_tagging.py

@@ -9,7 +9,11 @@ from sqlalchemy import select
 from backend.app.celery_app import celery
 from backend.app.models import Artist, ImageProvenance, ImageRecord, Post, Source
 from backend.app.models.tag import image_tag
-from backend.app.services.wip_title import apply_wip_image_tags, resolve_wip_tag_id
+from backend.app.services.wip_title import (
+    WIP_TITLE_SOFT_SOURCE,
+    apply_wip_image_tags,
+    resolve_wip_tag_id,
+)
 
 pytestmark = pytest.mark.integration
 
@@ -104,3 +108,14 @@ def test_backfill_tags_only_wip_titled_posts(db_sync):
 
     # Idempotent: a second sweep finds the tag already present and applies nothing.
     assert backfill_wip_title_tags.apply().get() == 0
+
+
+def test_apply_soft_source_stamps_wip_title_soft(db_sync):
+    # The soft tier (#1474) stamps a distinct provisional source.
+    tag_id = resolve_wip_tag_id(db_sync)
+    rec = _img(db_sync)
+    db_sync.commit()
+    assert apply_wip_image_tags(
+        db_sync, [rec.id], tag_id, source=WIP_TITLE_SOFT_SOURCE
+    ) == 1
+    assert _wip_source(db_sync, rec.id, tag_id) == "wip_title_soft"