Merge pull request 'Release v26.05.25.7 — FC-Cleanup tab + UniqueViolation fix + error modal + extension install fix' (#22) from dev into main

_apply_sidecar is additive: it find-or-creates Post/Source/ImageProvenance
and sets primary_post_id NULL-only, so any IR-migration provenance on the
existing row survives untouched and the new GS sidecar adds a second
ImageProvenance pointing at the freshly-created Post.

Wrapped in try/except so a malformed sidecar can't unwind the file-swap
commit — the file replacement is the critical operation.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

.forgejo/workflows/build.yml

@@ -6,18 +6,197 @@ on:
 
 # Requires repo secret RELEASE_TOKEN — a Forgejo PAT with scopes:
 #   - write:package, read:package  (for docker push to git.fabledsword.com)
-#   - write:release                (for future release-cutting workflows)
+#   - write:release                (for ext-<version> release asset cache)
 #   - write:issue                  (for future issue-management automation)
 # The injected GITHUB_TOKEN cannot be used — it lacks write:package.
 
 jobs:
-  build-web:
+  # Sign-or-fetch-from-cache: signs the extension via AMO if no ext-<version>
+  # Forgejo release exists yet, otherwise downloads the cached signed XPI.
+  # Result is uploaded as an Actions artifact for build-web to consume.
+  #
+  # Why this lives in build.yml (not a separate workflow): the merge-commit's
+  # docker image tagged `:latest` MUST carry the XPI. A separate sign workflow
+  # racing build.yml leaves `:latest` without the XPI for ~5min (until the
+  # commit-back triggers another build). Inline ordering eliminates the race.
+  # Cache strategy: Forgejo Release Assets — picked 2026-05-25 over Generic
+  # Packages (cleaner API surface) and commit-back-to-side-branch (no extra
+  # branch to manage). AMO blocks re-signing the same version (returns 409),
+  # so signing is intentionally one-shot per version bump.
+  sign-extension:
+    if: github.ref == 'refs/heads/main'
     runs-on: python-ci
     container:
       image: git.fabledsword.com/bvandeusen/ci-python:3.14
     steps:
       - uses: actions/checkout@v4
 
+      - name: Resolve extension version
+        id: extver
+        run: |
+          VERSION=$(grep -E '"version"' extension/package.json | head -1 | sed -E 's/.*"version"[[:space:]]*:[[:space:]]*"([^"]+)".*/\1/')
+          echo "version=$VERSION" >> "$GITHUB_OUTPUT"
+          echo "Resolved extension version: $VERSION"
+
+      - name: Check Forgejo release-asset cache
+        id: cache
+        env:
+          TOKEN: ${{ secrets.RELEASE_TOKEN }}
+        run: |
+          set -eu
+          VERSION=${{ steps.extver.outputs.version }}
+          STATUS=$(curl -s -o release.json -w "%{http_code}" \
+            -H "Authorization: token $TOKEN" \
+            "https://git.fabledsword.com/api/v1/repos/bvandeusen/FabledCurator/releases/tags/ext-$VERSION" || echo 000)
+          echo "Tag lookup HTTP status: $STATUS"
+          # JSON parsing via python (ci-python:3.14 has stdlib json; jq is
+          # not in the image and adding it per ci-requirements.md is not
+          # warranted for a single consumer — operator-flagged 2026-05-26
+          # after a sign job failed with `jq: not found`).
+          if [ "$STATUS" = "200" ]; then
+            ASSET_ID=$(python3 -c "import json; r=json.load(open('release.json')); xpis=[a for a in r.get('assets', []) if a.get('name','').endswith('.xpi')]; print(xpis[0]['id'] if xpis else '')")
+            if [ -n "$ASSET_ID" ]; then
+              echo "cached=true" >> "$GITHUB_OUTPUT"
+              echo "asset_id=$ASSET_ID" >> "$GITHUB_OUTPUT"
+              echo "Cached XPI exists at ext-$VERSION (asset id $ASSET_ID); skipping AMO sign"
+            else
+              echo "cached=false" >> "$GITHUB_OUTPUT"
+              echo "Release ext-$VERSION exists but has no .xpi asset; will re-sign + re-upload"
+            fi
+          else
+            echo "cached=false" >> "$GITHUB_OUTPUT"
+            echo "No release named ext-$VERSION; will sign via AMO and upload"
+          fi
+
+      # No "download cached XPI in sign-extension" step: build-web
+      # fetches directly from the Forgejo ext-<version> release asset
+      # (removed 2026-05-26 alongside the actions/upload-artifact
+      # removal — sign-extension's job is just to ensure the cache
+      # exists on Forgejo; the build-web side reads it independently).
+
+      - name: Sign via AMO (cache miss)
+        if: steps.cache.outputs.cached != 'true'
+        run: |
+          cd extension && npm install --no-save --no-audit --no-fund && npm run sign
+        env:
+          WEB_EXT_API_KEY: ${{ secrets.MOZILLA_AMO_JWT_KEY }}
+          WEB_EXT_API_SECRET: ${{ secrets.MOZILLA_AMO_JWT_SECRET }}
+
+      - name: Upload signed XPI to ext-<version> release (cache miss)
+        if: steps.cache.outputs.cached != 'true'
+        env:
+          TOKEN: ${{ secrets.RELEASE_TOKEN }}
+        run: |
+          set -eux
+          VERSION=${{ steps.extver.outputs.version }}
+          # AMO renames signed XPIs with its internal addon-id-safe-string;
+          # canonicalize to fabledcurator-<version>.xpi so the FC server's
+          # whitelist (backend/app/frontend.py expects 'fabledcurator-*.xpi')
+          # keeps working.
+          SIGNED=$(ls extension/web-ext-artifacts/*.xpi | head -1)
+          XPI="extension/web-ext-artifacts/fabledcurator-$VERSION.xpi"
+          cp "$SIGNED" "$XPI"
+          # Find-or-create the ext-<version> release. Track whether WE
+          # created it so an upload failure below can roll back (don't
+          # leave an empty release tombstone that the next run's
+          # cache-check mistakes for a partial-failure state).
+          STATUS=$(curl -s -o release.json -w "%{http_code}" \
+            -H "Authorization: token $TOKEN" \
+            "https://git.fabledsword.com/api/v1/repos/bvandeusen/FabledCurator/releases/tags/ext-$VERSION" || echo 000)
+          if [ "$STATUS" = "200" ]; then
+            CREATED_BY_US=false
+          else
+            curl -s -X POST -H "Authorization: token $TOKEN" -H "Content-Type: application/json" \
+              -d "{\"tag_name\":\"ext-$VERSION\",\"name\":\"Extension $VERSION (signed XPI cache)\",\"body\":\"Internal cache for the signed XPI consumed by build.yml's build-web job. Not a user-facing FC release.\",\"target_commitish\":\"main\"}" \
+              -o release.json \
+              "https://git.fabledsword.com/api/v1/repos/bvandeusen/FabledCurator/releases"
+            CREATED_BY_US=true
+          fi
+          RELEASE_ID=$(python3 -c "import json; print(json.load(open('release.json'))['id'])")
+          test -n "$RELEASE_ID"
+          # Rollback-on-failure: if the asset upload fails AND we just
+          # created the release in this run, delete it. Prevents an empty
+          # ext-<version> release from poisoning the next workflow run
+          # (operator-flagged 2026-05-26 — without rollback the next run
+          # saw 'release exists, no asset → cache miss → sign' which AMO
+          # then rejected with 409 'Version already exists').
+          rollback_if_we_created() {
+            if [ "$CREATED_BY_US" = "true" ]; then
+              echo "Rolling back: deleting just-created release $RELEASE_ID"
+              curl -s -X DELETE -H "Authorization: token $TOKEN" \
+                "https://git.fabledsword.com/api/v1/repos/bvandeusen/FabledCurator/releases/$RELEASE_ID" || true
+              curl -s -X DELETE -H "Authorization: token $TOKEN" \
+                "https://git.fabledsword.com/api/v1/repos/bvandeusen/FabledCurator/tags/ext-$VERSION" || true
+            fi
+          }
+          trap 'rollback_if_we_created' EXIT
+          HTTP_CODE=$(curl -s -X POST -H "Authorization: token $TOKEN" \
+            -F "attachment=@$XPI" \
+            -o /dev/null -w "%{http_code}" \
+            "https://git.fabledsword.com/api/v1/repos/bvandeusen/FabledCurator/releases/$RELEASE_ID/assets?name=fabledcurator-$VERSION.xpi")
+          if [ "$HTTP_CODE" != "201" ] && [ "$HTTP_CODE" != "200" ]; then
+            echo "Asset upload failed with HTTP $HTTP_CODE"
+            exit 1
+          fi
+          # Upload succeeded — clear the rollback trap.
+          trap - EXIT
+          echo "Uploaded fabledcurator-$VERSION.xpi to ext-$VERSION release"
+
+      # No actions/upload-artifact step: Forgejo Actions (and our
+       # act_runner) doesn't support upload-artifact@v4+ (GHES limitation
+       # surfaced 2026-05-26). Instead build-web reads the signed XPI
+       # straight from the ext-<version> Forgejo release we just uploaded
+       # to. Same source of truth; no double-store.
+
+  build-web:
+    needs: [sign-extension]
+    # sign-extension is main-only; on dev it's skipped, build-web still runs.
+    if: always() && (needs.sign-extension.result == 'success' || needs.sign-extension.result == 'skipped')
+    runs-on: python-ci
+    container:
+      image: git.fabledsword.com/bvandeusen/ci-python:3.14
+    steps:
+      - uses: actions/checkout@v4
+
+      - name: Download signed XPI from Forgejo release asset (main only)
+        if: github.ref == 'refs/heads/main'
+        env:
+          TOKEN: ${{ secrets.RELEASE_TOKEN }}
+        run: |
+          set -eux
+          VERSION=$(grep -E '"version"' extension/package.json | head -1 | sed -E 's/.*"version"[[:space:]]*:[[:space:]]*"([^"]+)".*/\1/')
+          # Look up the ext-<version> release; extract the .xpi asset's
+          # browser_download_url (Forgejo's /releases/assets/<id> endpoint
+          # returns ASSET METADATA, not the binary blob — operator-flagged
+          # 2026-05-26: my prior code curl'd the metadata endpoint without
+          # -f and wrote the resulting 404-page-not-found text into
+          # fabledcurator-*.xpi, which Firefox then rejected as "corrupt").
+          # browser_download_url is the canonical binary endpoint and is
+          # also publicly accessible (no token needed) but we pass the
+          # token anyway for symmetry with private-repo support.
+          curl -sf -H "Authorization: token $TOKEN" \
+            "https://git.fabledsword.com/api/v1/repos/bvandeusen/FabledCurator/releases/tags/ext-$VERSION" \
+            -o release.json
+          DOWNLOAD_URL=$(python3 -c "import json; r=json.load(open('release.json')); xpis=[a for a in r.get('assets', []) if a.get('name','').endswith('.xpi')]; print(xpis[0]['browser_download_url'])")
+          test -n "$DOWNLOAD_URL"
+          echo "Downloading XPI from: $DOWNLOAD_URL"
+          mkdir -p frontend/public/extension
+          DEST="frontend/public/extension/fabledcurator-$VERSION.xpi"
+          # -f = fail on HTTP error (prevents silent corruption like the
+          # 2026-05-26 incident); -L = follow redirects.
+          curl -sfL -H "Authorization: token $TOKEN" -o "$DEST" "$DOWNLOAD_URL"
+          # Sanity check: the binary should start with the ZIP magic (PK\x03\x04).
+          # If it's anything else, the next docker build will ship a corrupt XPI.
+          MAGIC=$(head -c 2 "$DEST" | od -An -c | tr -d ' \n')
+          if [ "$MAGIC" != "PK" ]; then
+            echo "ERROR: downloaded XPI does not start with ZIP magic 'PK' (got '$MAGIC')"
+            echo "File contents preview:"
+            head -c 200 "$DEST"
+            exit 1
+          fi
+          cp "$DEST" "frontend/public/extension/fabledcurator-latest.xpi"
+          ls -la frontend/public/extension/
+
       - name: Determine tag
         id: tag
         run: |

.forgejo/workflows/ci.yml

@@ -24,11 +24,26 @@ jobs:
     steps:
       - uses: actions/checkout@v4
 
+      - name: Cache pip wheels
+        uses: actions/cache@v4
+        with:
+          path: ~/.cache/pip
+          key: pip-${{ runner.os }}-py314-${{ hashFiles('requirements.txt') }}
+          restore-keys: |
+            pip-${{ runner.os }}-py314-
+
       - name: Install Python deps
         # ruff is pre-installed in the ci-python image (see CI-Runner/CI-python/
         # Dockerfile's RUFF_VERSION). Per FabledRulebook ci-runners.md, toolchain
         # versions live on the runner image, not here.
-        run: pip install -r requirements.txt pytest pytest-asyncio
+        # uv: 5-10x faster wheel resolve than pip for cold caches.
+        # Falls back to pip install on uv-missing runners (older images).
+        run: |
+          if command -v uv >/dev/null 2>&1; then
+            uv pip install --system -r requirements.txt pytest pytest-asyncio
+          else
+            pip install -r requirements.txt pytest pytest-asyncio
+          fi
 
       - name: Ruff lint
         run: ruff check backend/ tests/ alembic/
@@ -57,17 +72,28 @@ jobs:
       - run: npm run test:unit
       - run: npm run build
 
-  integration:
-    # This act_runner (swarm-runner v0.6.1) puts service containers on the
-    # default bridge with NO service-name DNS, and publishing fixed host
-    # ports collides with the operator's running docker-compose dev stack on
-    # the same shared daemon. Workaround: publish NO host ports, and reach
-    # each service by its bridge IP — discovered at runtime via the mounted
-    # docker socket (the ci-python image ships /usr/bin/docker). Default-bridge
-    # containers can talk by IP (only embedded DNS is missing), so IP
-    # addressing is reliable here. Everything runs in ONE step so resolved
-    # values don't depend on cross-step env passing. Pattern documented in
-    # FabledRulebook/forgejo.md "CI philosophy".
+  # Integration suite split into THREE parallel shards (2026-05-25, runner
+  # capacity bumped 2→6). Each shard gets its own Postgres + Redis service
+  # set and runs alembic + a disjoint subset of integration tests. Shards
+  # share no DB state, so the autouse TRUNCATE fixture in tests/conftest.py
+  # stays single-threaded per shard but multiple shards run in parallel
+  # wall-clock. Approximate split — rebalance once --durations=15 output
+  # reveals which shard is the long pole.
+  #
+  # Each shard's docker-ps filter uses its own unique job name to scope
+  # service-container resolution. act_runner appears to strip underscores
+  # from job names when building container labels — `int_api` yielded
+  # zero matches on 2026-05-25 — so shards use no-separator names
+  # (`intapi`, `intimp`, `intcore`) instead. Each step prints
+  # `docker ps -a` first so a future naming-convention shift surfaces in
+  # the log without another guess-and-push cycle.
+  #
+  # Pre-baking requirements.txt into ci-python:3.14 is intentionally NOT
+  # done — per ci-requirements.md, FC is the only Python consumer of that
+  # image and the CI-Runner project's "add deps to image when used by >1
+  # project" rule keeps the install per-job.
+
+  intapi:
     runs-on: python-ci
     container:
       image: git.fabledsword.com/bvandeusen/ci-python:3.14
@@ -98,15 +124,21 @@ jobs:
           --health-retries 10
     steps:
       - uses: actions/checkout@v4
-
-      - name: Integration suite (resolve service IPs, migrate, test)
+      - name: Cache pip wheels
+        uses: actions/cache@v4
+        with:
+          path: ~/.cache/pip
+          key: pip-${{ runner.os }}-py314-${{ hashFiles('requirements.txt') }}
+          restore-keys: |
+            pip-${{ runner.os }}-py314-
+      - name: API integration shard (resolve service IPs, migrate, test)
         run: |
           set -eux
-          # Scope to THIS job's service containers (act_runner names them
-          # ...JOB-integration...); the operator's compose stack uses the
-          # same images but different names, so it won't match.
-          PG=$(docker ps --filter "name=JOB-integration" --filter "ancestor=pgvector/pgvector:pg16" -q | head -n1)
-          RD=$(docker ps --filter "name=JOB-integration" --filter "ancestor=redis:7-alpine" -q | head -n1)
+          echo "=== container landscape (diagnostic for filter scoping) ==="
+          docker ps -a --format '{{.ID}} {{.Image}} -> {{.Names}}'
+          echo "=== end landscape ==="
+          PG=$(docker ps --filter "name=intapi" --filter "ancestor=pgvector/pgvector:pg16" -q | head -n1)
+          RD=$(docker ps --filter "name=intapi" --filter "ancestor=redis:7-alpine" -q | head -n1)
           test -n "$PG" && test -n "$RD"
           PG_IP=$(docker inspect -f '{{range .NetworkSettings.Networks}}{{.IPAddress}}{{end}}' "$PG")
           RD_IP=$(docker inspect -f '{{range .NetworkSettings.Networks}}{{.IPAddress}}{{end}}' "$RD")
@@ -114,11 +146,153 @@ jobs:
           export DB_HOST="$PG_IP"
           export CELERY_BROKER_URL="redis://$RD_IP:6379/0"
           export CELERY_RESULT_BACKEND="redis://$RD_IP:6379/0"
-          # Wait for Postgres to accept TCP (bash /dev/tcp; no extra tools).
           for i in $(seq 1 60); do
             (echo > "/dev/tcp/$PG_IP/5432") >/dev/null 2>&1 && break
             sleep 2
           done
-          pip install -r requirements.txt pytest pytest-asyncio
+          if command -v uv >/dev/null 2>&1; then
+            uv pip install --system -r requirements.txt pytest pytest-asyncio
+          else
+            pip install -r requirements.txt pytest pytest-asyncio
+          fi
           alembic upgrade head
-          pytest tests/ -v -m integration
+          pytest tests/test_api_*.py -v -m integration --durations=15
+
+  intimp:
+    runs-on: python-ci
+    container:
+      image: git.fabledsword.com/bvandeusen/ci-python:3.14
+    env:
+      DB_USER: fabledcurator
+      DB_PASSWORD: ci_integration
+      DB_PORT: "5432"
+      DB_NAME: fabledcurator_test
+      SECRET_KEY: ci_integration_placeholder
+    services:
+      postgres:
+        image: pgvector/pgvector:pg16
+        env:
+          POSTGRES_USER: fabledcurator
+          POSTGRES_PASSWORD: ci_integration
+          POSTGRES_DB: fabledcurator_test
+        options: >-
+          --health-cmd "pg_isready -U fabledcurator"
+          --health-interval 10s
+          --health-timeout 5s
+          --health-retries 10
+      redis:
+        image: redis:7-alpine
+        options: >-
+          --health-cmd "redis-cli ping"
+          --health-interval 10s
+          --health-timeout 5s
+          --health-retries 10
+    steps:
+      - uses: actions/checkout@v4
+      - name: Cache pip wheels
+        uses: actions/cache@v4
+        with:
+          path: ~/.cache/pip
+          key: pip-${{ runner.os }}-py314-${{ hashFiles('requirements.txt') }}
+          restore-keys: |
+            pip-${{ runner.os }}-py314-
+      - name: Importer integration shard (resolve service IPs, migrate, test)
+        run: |
+          set -eux
+          echo "=== container landscape (diagnostic for filter scoping) ==="
+          docker ps -a --format '{{.ID}} {{.Image}} -> {{.Names}}'
+          echo "=== end landscape ==="
+          PG=$(docker ps --filter "name=intimp" --filter "ancestor=pgvector/pgvector:pg16" -q | head -n1)
+          RD=$(docker ps --filter "name=intimp" --filter "ancestor=redis:7-alpine" -q | head -n1)
+          test -n "$PG" && test -n "$RD"
+          PG_IP=$(docker inspect -f '{{range .NetworkSettings.Networks}}{{.IPAddress}}{{end}}' "$PG")
+          RD_IP=$(docker inspect -f '{{range .NetworkSettings.Networks}}{{.IPAddress}}{{end}}' "$RD")
+          test -n "$PG_IP" && test -n "$RD_IP"
+          export DB_HOST="$PG_IP"
+          export CELERY_BROKER_URL="redis://$RD_IP:6379/0"
+          export CELERY_RESULT_BACKEND="redis://$RD_IP:6379/0"
+          for i in $(seq 1 60); do
+            (echo > "/dev/tcp/$PG_IP/5432") >/dev/null 2>&1 && break
+            sleep 2
+          done
+          if command -v uv >/dev/null 2>&1; then
+            uv pip install --system -r requirements.txt pytest pytest-asyncio
+          else
+            pip install -r requirements.txt pytest pytest-asyncio
+          fi
+          alembic upgrade head
+          pytest tests/test_importer*.py tests/test_import_*.py tests/test_migration_*.py tests/test_phash_*.py tests/test_sidecar_*.py tests/test_scan_*.py tests/test_archive_extractor.py tests/test_backfill_phash.py -v -m integration --durations=15
+
+  intcore:
+    runs-on: python-ci
+    container:
+      image: git.fabledsword.com/bvandeusen/ci-python:3.14
+    env:
+      DB_USER: fabledcurator
+      DB_PASSWORD: ci_integration
+      DB_PORT: "5432"
+      DB_NAME: fabledcurator_test
+      SECRET_KEY: ci_integration_placeholder
+    services:
+      postgres:
+        image: pgvector/pgvector:pg16
+        env:
+          POSTGRES_USER: fabledcurator
+          POSTGRES_PASSWORD: ci_integration
+          POSTGRES_DB: fabledcurator_test
+        options: >-
+          --health-cmd "pg_isready -U fabledcurator"
+          --health-interval 10s
+          --health-timeout 5s
+          --health-retries 10
+      redis:
+        image: redis:7-alpine
+        options: >-
+          --health-cmd "redis-cli ping"
+          --health-interval 10s
+          --health-timeout 5s
+          --health-retries 10
+    steps:
+      - uses: actions/checkout@v4
+      - name: Cache pip wheels
+        uses: actions/cache@v4
+        with:
+          path: ~/.cache/pip
+          key: pip-${{ runner.os }}-py314-${{ hashFiles('requirements.txt') }}
+          restore-keys: |
+            pip-${{ runner.os }}-py314-
+      - name: Core integration shard (everything not api / importer / migration / phash / sidecar / scan / archive / backfill)
+        run: |
+          set -eux
+          echo "=== container landscape (diagnostic for filter scoping) ==="
+          docker ps -a --format '{{.ID}} {{.Image}} -> {{.Names}}'
+          echo "=== end landscape ==="
+          PG=$(docker ps --filter "name=intcore" --filter "ancestor=pgvector/pgvector:pg16" -q | head -n1)
+          RD=$(docker ps --filter "name=intcore" --filter "ancestor=redis:7-alpine" -q | head -n1)
+          test -n "$PG" && test -n "$RD"
+          PG_IP=$(docker inspect -f '{{range .NetworkSettings.Networks}}{{.IPAddress}}{{end}}' "$PG")
+          RD_IP=$(docker inspect -f '{{range .NetworkSettings.Networks}}{{.IPAddress}}{{end}}' "$RD")
+          test -n "$PG_IP" && test -n "$RD_IP"
+          export DB_HOST="$PG_IP"
+          export CELERY_BROKER_URL="redis://$RD_IP:6379/0"
+          export CELERY_RESULT_BACKEND="redis://$RD_IP:6379/0"
+          for i in $(seq 1 60); do
+            (echo > "/dev/tcp/$PG_IP/5432") >/dev/null 2>&1 && break
+            sleep 2
+          done
+          if command -v uv >/dev/null 2>&1; then
+            uv pip install --system -r requirements.txt pytest pytest-asyncio
+          else
+            pip install -r requirements.txt pytest pytest-asyncio
+          fi
+          alembic upgrade head
+          pytest tests/ -v -m integration --durations=15 \
+            --ignore-glob='tests/test_api_*.py' \
+            --ignore-glob='tests/test_importer*.py' \
+            --ignore-glob='tests/test_import_*.py' \
+            --ignore-glob='tests/test_migration_*.py' \
+            --ignore-glob='tests/test_phash_*.py' \
+            --ignore-glob='tests/test_sidecar_*.py' \
+            --ignore-glob='tests/test_scan_*.py' \
+            --ignore-glob='tests/test_archive_extractor.py' \
+            --ignore-glob='tests/test_backfill_phash.py'

.forgejo/workflows/extension.yml

@@ -1,8 +1,19 @@
 name: extension
+# Lint-only workflow. The sign-and-publish dance moved into build.yml's
+# `sign-extension` job (2026-05-25) — `:latest` now always bundles the XPI
+# because sign-extension runs as a build-web dependency in the SAME workflow,
+# eliminating the prior race between build.yml and a separate extension.yml.
+# Signed XPIs are cached in Forgejo Release Assets named `ext-<version>`.
 on:
   push:
     branches: [dev, main]
-    paths: ['extension/**']
+    paths:
+      - 'extension/**'
+      - '.forgejo/workflows/extension.yml'
+  pull_request:
+    branches: [main]
+    paths:
+      - 'extension/**'
   workflow_dispatch:
 
 jobs:
@@ -16,44 +27,3 @@ jobs:
         run: cd extension && npm install --no-save --no-audit --no-fund
       - name: Lint
         run: cd extension && npm run lint
-
-  sign-and-publish:
-    needs: lint
-    if: github.ref == 'refs/heads/main'
-    runs-on: python-ci
-    container:
-      image: node:22-bookworm-slim
-    steps:
-      - uses: actions/checkout@v4
-        with:
-          token: ${{ secrets.RELEASE_TOKEN }}
-      - name: Install web-ext + git
-        run: |
-          apt-get update && apt-get install -y --no-install-recommends git ca-certificates
-          cd extension && npm install --no-save --no-audit --no-fund
-      - name: Sign XPI
-        run: cd extension && npm run sign
-        env:
-          WEB_EXT_API_KEY: ${{ secrets.MOZILLA_AMO_JWT_KEY }}
-          WEB_EXT_API_SECRET: ${{ secrets.MOZILLA_AMO_JWT_SECRET }}
-      - name: Commit signed XPI to frontend/public/extension/
-        run: |
-          set -e
-          XPI=$(ls extension/web-ext-artifacts/fabledcurator-*.xpi | head -1)
-          if [ -z "$XPI" ]; then
-            echo "No XPI produced by web-ext sign — exiting"
-            exit 1
-          fi
-          mkdir -p frontend/public/extension
-          cp "$XPI" frontend/public/extension/
-          # Also copy as -latest.xpi so the FC server can serve a stable URL.
-          cp "$XPI" "frontend/public/extension/fabledcurator-latest.xpi"
-          git config user.name  "FC extension CI"
-          git config user.email "noreply@fabledsword.com"
-          git add frontend/public/extension/
-          if git diff --cached --quiet; then
-            echo "No changes to commit"
-          else
-            git commit -m "ext: publish signed XPI $(basename $XPI)"
-            git push origin HEAD:main
-          fi

alembic/versions/0019_import_batch_refreshed.py

@@ -0,0 +1,38 @@
+"""import_batch.refreshed counter for deep-scan sidecar re-application
+
+Revision ID: 0019
+Revises: 0018
+Create Date: 2026-05-25
+
+Adds a `refreshed` counter to `import_batch`, mirroring the existing
+`imported`/`skipped`/`failed`/`attachments` columns. Deep scan now
+re-applies sidecar metadata to already-imported files (the IR feature
+that didn't make the FC port the first time); a "refreshed" outcome
+increments this counter so the UI can surface "X new, Y refreshed"
+instead of the misleading "Scan complete — no new files" message.
+
+server_default=0 backfills existing rows in place — no UPDATE needed.
+"""
+from typing import Sequence, Union
+
+import sqlalchemy as sa
+from alembic import op
+
+revision: str = "0019"
+down_revision: Union[str, None] = "0018"
+branch_labels: Union[str, Sequence[str], None] = None
+depends_on: Union[str, Sequence[str], None] = None
+
+
+def upgrade() -> None:
+    op.add_column(
+        "import_batch",
+        sa.Column(
+            "refreshed", sa.Integer(),
+            nullable=False, server_default=sa.text("0"),
+        ),
+    )
+
+
+def downgrade() -> None:
+    op.drop_column("import_batch", "refreshed")

alembic/versions/0020_library_audit_run.py

@@ -0,0 +1,65 @@
+"""fc-cleanup: library_audit_run table for async transparency/single_color audits
+
+Revision ID: 0020
+Revises: 0019
+Create Date: 2026-05-26
+
+The table backs the async audit lifecycle: rule + params snapshot, status
+state machine ('running' → 'ready' → 'applied'/'cancelled'/'error'), and
+the matched_ids JSONB array that the apply step deletes. Capped at 50k IDs
+per row by the scan task (oversize = rule too aggressive, operator narrows
+before re-running).
+"""
+from typing import Sequence, Union
+
+import sqlalchemy as sa
+from alembic import op
+from sqlalchemy.dialects import postgresql
+
+revision: str = "0020"
+down_revision: Union[str, None] = "0019"
+branch_labels: Union[str, Sequence[str], None] = None
+depends_on: Union[str, Sequence[str], None] = None
+
+
+def upgrade() -> None:
+    op.create_table(
+        "library_audit_run",
+        sa.Column("id", sa.Integer(), primary_key=True),
+        sa.Column("rule", sa.String(32), nullable=False),
+        sa.Column("params", postgresql.JSONB(astext_type=sa.Text()), nullable=False),
+        sa.Column(
+            "status", sa.String(16),
+            nullable=False, server_default="running",
+        ),
+        sa.Column(
+            "started_at", sa.DateTime(timezone=True),
+            nullable=False, server_default=sa.func.now(),
+        ),
+        sa.Column("finished_at", sa.DateTime(timezone=True), nullable=True),
+        sa.Column(
+            "scanned_count", sa.Integer(),
+            nullable=False, server_default="0",
+        ),
+        sa.Column(
+            "matched_count", sa.Integer(),
+            nullable=False, server_default="0",
+        ),
+        sa.Column(
+            "matched_ids", postgresql.JSONB(astext_type=sa.Text()),
+            nullable=False, server_default=sa.text("'[]'::jsonb"),
+        ),
+        sa.Column("error", sa.Text(), nullable=True),
+    )
+    op.create_index(
+        "ix_library_audit_run_rule", "library_audit_run", ["rule"],
+    )
+    op.create_index(
+        "ix_library_audit_run_status", "library_audit_run", ["status"],
+    )
+
+
+def downgrade() -> None:
+    op.drop_index("ix_library_audit_run_status", table_name="library_audit_run")
+    op.drop_index("ix_library_audit_run_rule", table_name="library_audit_run")
+    op.drop_table("library_audit_run")

backend/app/api/__init__.py

@@ -20,6 +20,7 @@ def all_blueprints() -> list[Blueprint]:
     from .artist import artist_bp
     from .artists import artists_bp
     from .attachments import attachments_bp
+    from .cleanup import cleanup_bp
     from .credentials import credentials_bp
     from .downloads import downloads_bp
     from .extension import extension_bp
@@ -50,6 +51,7 @@ def all_blueprints() -> list[Blueprint]:
         system_activity_bp,
         system_backup_bp,
         admin_bp,
+        cleanup_bp,
         import_admin_bp,
         migrate_bp,
         suggestions_bp,

backend/app/api/cleanup.py

@@ -0,0 +1,193 @@
+"""FC-Cleanup: /api/cleanup/* — retroactive enforcement of import filters.
+
+Endpoints:
+  POST /min-dimension/preview     synchronous SQL audit
+  POST /min-dimension/delete      synchronous SQL delete (Tier-C token)
+  POST /audit                      async transparency / single_color start
+  GET  /audit                      list recent audit_run rows
+  GET  /audit/<id>                 single audit_run row
+  POST /audit/<id>/apply           apply matched_ids deletes (Tier-C token)
+  POST /audit/<id>/cancel          flip running audit to cancelled
+
+Unused-tags retroactive prune intentionally NOT in this namespace —
+TagMaintenanceCard (Maintenance tab → moved to Cleanup tab in v26.05.25.7)
+uses the existing /api/admin/tags/prune-unused endpoint via the admin
+store. No duplicate route here.
+
+Confirm-token format matches modal/DestructiveConfirmModal.vue convention:
+  `delete-min-dim-<sha8(w,h)>` for min-dim delete
+  `delete-audit-<id>` for audit apply
+(Modal hardcodes action ∈ {'restore', 'delete'}; "apply audit" is semantically a delete of the matched images, so we use `delete-audit-<id>`.)
+"""
+
+from __future__ import annotations
+
+import hashlib
+from pathlib import Path
+
+from quart import Blueprint, jsonify, request
+from sqlalchemy import select
+
+from ..extensions import get_session
+from ..models import LibraryAuditRun
+from ..services import cleanup_service
+
+cleanup_bp = Blueprint("cleanup", __name__, url_prefix="/api/cleanup")
+
+IMAGES_ROOT = Path("/images")
+
+
+def _bad(error: str, *, status: int = 400, **extra):
+    body = {"error": error}
+    body.update(extra)
+    return jsonify(body), status
+
+
+def _min_dim_token(min_w: int, min_h: int) -> str:
+    # SHA-256 (not MD5) — Web Crypto's subtle.digest rejects MD5; both
+    # sides use SHA-256 truncated to 8 hex chars.
+    canon = f"{min_w}x{min_h}"
+    return f"delete-min-dim-{hashlib.sha256(canon.encode()).hexdigest()[:8]}"
+
+
+def _serialize_audit_run(audit: LibraryAuditRun) -> dict:
+    return {
+        "id": audit.id,
+        "rule": audit.rule,
+        "params": audit.params,
+        "status": audit.status,
+        "started_at": audit.started_at.isoformat() if audit.started_at else None,
+        "finished_at": audit.finished_at.isoformat() if audit.finished_at else None,
+        "scanned_count": audit.scanned_count,
+        "matched_count": audit.matched_count,
+        "matched_ids": audit.matched_ids,
+        "error": audit.error,
+    }
+
+
+@cleanup_bp.route("/min-dimension/preview", methods=["POST"])
+async def min_dim_preview():
+    body = await request.get_json(silent=True) or {}
+    try:
+        min_w = int(body.get("min_width", 0))
+        min_h = int(body.get("min_height", 0))
+    except (TypeError, ValueError):
+        return _bad("invalid_dimensions")
+    if min_w < 0 or min_h < 0:
+        return _bad("invalid_dimensions")
+    async with get_session() as session:
+        projection = await session.run_sync(
+            lambda s: cleanup_service.project_min_dimension_violations(
+                s, min_width=min_w, min_height=min_h,
+            )
+        )
+    return jsonify(projection)
+
+
+@cleanup_bp.route("/min-dimension/delete", methods=["POST"])
+async def min_dim_delete():
+    body = await request.get_json(silent=True) or {}
+    try:
+        min_w = int(body.get("min_width", 0))
+        min_h = int(body.get("min_height", 0))
+    except (TypeError, ValueError):
+        return _bad("invalid_dimensions")
+    if min_w < 0 or min_h < 0:
+        return _bad("invalid_dimensions")
+    supplied = body.get("confirm", "")
+    expected = _min_dim_token(min_w, min_h)
+    if supplied != expected:
+        return _bad("confirm_mismatch", expected=expected)
+    async with get_session() as session:
+        deleted = await session.run_sync(
+            lambda s: cleanup_service.delete_min_dimension_violations(
+                s, min_width=min_w, min_height=min_h, images_root=IMAGES_ROOT,
+            )
+        )
+        await session.commit()
+    return jsonify({"deleted": deleted})
+
+
+@cleanup_bp.route("/audit", methods=["POST"])
+async def audit_create():
+    body = await request.get_json(silent=True) or {}
+    rule = body.get("rule")
+    params = body.get("params") or {}
+    if rule not in ("transparency", "single_color"):
+        return _bad("invalid_rule")
+    if not isinstance(params, dict):
+        return _bad("invalid_params")
+    async with get_session() as session:
+        try:
+            audit_id = await session.run_sync(
+                lambda s: cleanup_service.start_audit_run(
+                    s, rule=rule, params=params,
+                )
+            )
+        except cleanup_service.AuditAlreadyRunning as running_id:
+            return _bad(
+                "audit_already_running", status=409,
+                running_id=int(str(running_id)),
+            )
+        except ValueError as exc:
+            return _bad(str(exc))
+        await session.commit()
+    return jsonify({"audit_id": audit_id, "status": "running"}), 202
+
+
+@cleanup_bp.route("/audit/<int:audit_id>", methods=["GET"])
+async def audit_get(audit_id: int):
+    async with get_session() as session:
+        audit = (await session.execute(
+            select(LibraryAuditRun).where(LibraryAuditRun.id == audit_id)
+        )).scalar_one_or_none()
+        if audit is None:
+            return _bad("not_found", status=404)
+        return jsonify(_serialize_audit_run(audit))
+
+
+@cleanup_bp.route("/audit", methods=["GET"])
+async def audit_history():
+    try:
+        limit = min(int(request.args.get("limit", "20")), 100)
+    except ValueError:
+        return _bad("invalid_limit")
+    async with get_session() as session:
+        rows = (await session.execute(
+            select(LibraryAuditRun)
+            .order_by(LibraryAuditRun.id.desc())
+            .limit(limit)
+        )).scalars().all()
+        return jsonify({"runs": [_serialize_audit_run(r) for r in rows]})
+
+
+@cleanup_bp.route("/audit/<int:audit_id>/apply", methods=["POST"])
+async def audit_apply(audit_id: int):
+    body = await request.get_json(silent=True) or {}
+    confirm = body.get("confirm", "")
+    async with get_session() as session:
+        try:
+            deleted = await session.run_sync(
+                lambda s: cleanup_service.apply_audit_run(
+                    s, audit_id=audit_id, confirm_token=confirm,
+                    images_root=IMAGES_ROOT,
+                )
+            )
+        except cleanup_service.AuditNotReady as exc:
+            return _bad("audit_not_ready", current_status=str(exc))
+        except cleanup_service.ConfirmTokenMismatch as exc:
+            return _bad("confirm_mismatch", expected=str(exc))
+        except ValueError as exc:
+            return _bad("not_found", status=404, detail=str(exc))
+        await session.commit()
+    return jsonify({"deleted": deleted})
+
+
+@cleanup_bp.route("/audit/<int:audit_id>/cancel", methods=["POST"])
+async def audit_cancel(audit_id: int):
+    async with get_session() as session:
+        await session.run_sync(
+            lambda s: cleanup_service.cancel_audit_run(s, audit_id=audit_id)
+        )
+        await session.commit()
+    return jsonify({"cancelled": True})

backend/app/api/extension.py

@@ -93,10 +93,20 @@ def _read_manifest_sync() -> dict | None:
     asyncio.to_thread (ASYNC240: no pathlib I/O in async functions)."""
     if not XPI_DIR.is_dir():
         return None
-    xpis = sorted(XPI_DIR.glob("fabledcurator-*.xpi"), key=lambda p: p.stat().st_mtime)
-    if not xpis:
+    # Exclude the `fabledcurator-latest.xpi` alias when picking the file to
+    # extract a version from — it's a copy of the latest versioned XPI,
+    # written at the same mtime by build.yml, and would otherwise tie or
+    # win the sort (operator-flagged 2026-05-26: UI displayed "v latest"
+    # because `_extract_version("fabledcurator-latest.xpi")` returns
+    # the literal "latest"). The alias still serves as `latest_url`.
+    versioned = [
+        p for p in XPI_DIR.glob("fabledcurator-*.xpi")
+        if p.name != "fabledcurator-latest.xpi"
+    ]
+    if not versioned:
         return None
-    latest = xpis[-1]
+    versioned.sort(key=lambda p: p.stat().st_mtime)
+    latest = versioned[-1]
     return {
         "installed": True,
         "version": _extract_version(latest.name),

backend/app/api/import_admin.py

@@ -47,10 +47,13 @@ async def status():
         if active:
             payload["active_batch"] = {
                 "id": active.id,
+                "source_path": active.source_path,
+                "scan_mode": active.scan_mode,
                 "total_files": active.total_files,
                 "imported": active.imported,
                 "skipped": active.skipped,
                 "failed": active.failed,
+                "refreshed": active.refreshed,
                 "started_at": active.started_at.isoformat(),
             }
         return jsonify(payload)

backend/app/celery_app.py

@@ -33,6 +33,7 @@ def make_celery() -> Celery:
             "backend.app.tasks.download",
             "backend.app.tasks.backup",
             "backend.app.tasks.admin",
+            "backend.app.tasks.library_audit",
         ],
     )
     app.conf.update(
@@ -47,6 +48,7 @@ def make_celery() -> Celery:
             "backend.app.tasks.migration.*": {"queue": "maintenance"},
             "backend.app.tasks.backup.*": {"queue": "maintenance"},
             "backend.app.tasks.admin.*": {"queue": "maintenance"},
+            "backend.app.tasks.library_audit.*": {"queue": "maintenance"},
         },
         # Heavy ML tasks need fair dispatch — see ImageRepo's precedent.
         task_acks_late=True,

backend/app/models/__init__.py

@@ -11,6 +11,7 @@ from .image_record import ImageRecord
 from .import_batch import ImportBatch
 from .import_settings import ImportSettings
 from .import_task import ImportTask
+from .library_audit_run import LibraryAuditRun
 from .migration_run import MigrationRun
 from .ml_settings import MLSettings
 from .post import Post
@@ -43,6 +44,7 @@ __all__ = [
     "ImportBatch",
     "ImportTask",
     "ImportSettings",
+    "LibraryAuditRun",
     "MLSettings",
     "MigrationRun",
     "TagAlias",

backend/app/models/import_batch.py

@@ -26,6 +26,10 @@ class ImportBatch(Base):
     skipped: Mapped[int] = mapped_column(Integer, nullable=False, default=0)
     failed: Mapped[int] = mapped_column(Integer, nullable=False, default=0)
     attachments: Mapped[int] = mapped_column(Integer, nullable=False, default=0)
+    # Deep-scan only: count of already-imported files whose sidecar metadata
+    # got re-applied this run (post/source/provenance upsert). Stays 0 on
+    # quick-scan batches. See `Importer.import_one(deep_scan=True)`.
+    refreshed: Mapped[int] = mapped_column(Integer, nullable=False, default=0)
 
     status: Mapped[str] = mapped_column(String(16), nullable=False, default="running", index=True)
     # running | complete | cancelled

backend/app/models/library_audit_run.py

@@ -0,0 +1,37 @@
+"""LibraryAuditRun — async transparency / single_color audit lifecycle.
+
+State machine: running → ready → applied / cancelled / error.
+matched_ids JSONB is appended-to by scan_library_for_rule; apply_audit_run
+reads it and routes through cleanup_service.delete_images.
+"""
+
+from datetime import datetime
+from typing import Any
+
+from sqlalchemy import DateTime, Integer, String, Text, func
+from sqlalchemy.dialects.postgresql import JSONB
+from sqlalchemy.orm import Mapped, mapped_column
+
+from .base import Base
+
+
+class LibraryAuditRun(Base):
+    __tablename__ = "library_audit_run"
+
+    id: Mapped[int] = mapped_column(Integer, primary_key=True)
+    rule: Mapped[str] = mapped_column(String(32), nullable=False, index=True)
+    params: Mapped[dict[str, Any]] = mapped_column(JSONB, nullable=False)
+    status: Mapped[str] = mapped_column(
+        String(16), nullable=False, default="running", index=True,
+    )
+    # running | ready | applied | cancelled | error
+    started_at: Mapped[datetime] = mapped_column(
+        DateTime(timezone=True), nullable=False, server_default=func.now(),
+    )
+    finished_at: Mapped[datetime | None] = mapped_column(
+        DateTime(timezone=True), nullable=True,
+    )
+    scanned_count: Mapped[int] = mapped_column(Integer, nullable=False, default=0)
+    matched_count: Mapped[int] = mapped_column(Integer, nullable=False, default=0)
+    matched_ids: Mapped[list[int]] = mapped_column(JSONB, nullable=False, default=list)
+    error: Mapped[str | None] = mapped_column(Text, nullable=True)

backend/app/services/audits/__init__.py

@@ -0,0 +1,6 @@
+"""Audit rule modules. Each module exposes evaluate(pil_image, **params) -> bool.
+
+The retroactive library-cleanup tab and (future) import-time filter logic
+both consume these. Importers should NOT inline rule logic going forward;
+add the rule here and call from both sides.
+"""

backend/app/services/audits/single_color.py

@@ -0,0 +1,53 @@
+"""Single-color audit: matches images where one color dominates beyond
+the threshold (within the given Euclidean RGB tolerance). The first
+canonical implementation — the import-side filter (SkipReason.single_color)
+was never wired; FC-Cleanup's audit module is the source of truth and a
+future spec can adopt it on the import path too.
+"""
+
+from PIL import Image
+
+_THUMB_SIZE = (64, 64)
+
+
+def evaluate(
+    pil_image,
+    *,
+    threshold: float,
+    tolerance: int,
+) -> bool:
+    """True iff the fraction of pixels within `tolerance` (Euclidean RGB
+    distance) of the dominant color exceeds `threshold`.
+
+    Downsamples to 64x64 for speed (~4ms regardless of source size).
+    Alpha channels are stripped; only RGB is considered. Animated images
+    use frame 0 (PIL's default after Image.open without seek).
+    """
+    im = pil_image
+    if im.mode == "RGBA":
+        im = im.convert("RGB")
+    elif im.mode not in ("RGB", "L"):
+        im = im.convert("RGB")
+    if im.size != _THUMB_SIZE:
+        im = im.resize(_THUMB_SIZE, Image.Resampling.BILINEAR)
+    pixels = list(im.getdata())
+    if not pixels:
+        return False
+    # Normalize L-mode pixels to RGB tuples for distance math.
+    if isinstance(pixels[0], int):
+        pixels = [(p, p, p) for p in pixels]
+    # Dominant color = mean RGB.
+    n = len(pixels)
+    sum_r = sum(p[0] for p in pixels)
+    sum_g = sum(p[1] for p in pixels)
+    sum_b = sum(p[2] for p in pixels)
+    dom = (sum_r / n, sum_g / n, sum_b / n)
+    tol_sq = tolerance * tolerance
+    within = 0
+    for r, g, b in pixels:
+        dr = r - dom[0]
+        dg = g - dom[1]
+        db = b - dom[2]
+        if dr * dr + dg * dg + db * db <= tol_sq:
+            within += 1
+    return (within / n) > threshold

backend/app/services/audits/transparency.py

@@ -0,0 +1,27 @@
+"""Transparency audit: matches images whose transparent-pixel fraction
+exceeds the threshold. Animated images short-circuit (skipped) to avoid
+the multi-frame PIL decode that hits Celery's hard time limit."""
+
+
+def evaluate(pil_image, *, threshold: float) -> bool:
+    """True iff the image's transparent-pixel fraction exceeds threshold.
+
+    False for non-alpha modes and animated images. Mirrors the import-side
+    Importer._transparency_pct logic so retroactive enforcement matches
+    prospective filtering.
+    """
+    if getattr(pil_image, "is_animated", False):
+        return False
+    if pil_image.mode not in ("RGBA", "LA") and not (
+        pil_image.mode == "P" and "transparency" in pil_image.info
+    ):
+        return False
+    im = pil_image
+    if im.mode != "RGBA":
+        im = im.convert("RGBA")
+    alpha = im.getchannel("A")
+    histogram = alpha.histogram()
+    transparent = histogram[0]
+    total = sum(histogram)
+    pct = transparent / total if total else 0.0
+    return pct > threshold

backend/app/services/cleanup_service.py

@@ -12,12 +12,14 @@ re-exports from this module and then delete the wrapper.
 """
 from __future__ import annotations
 
+from datetime import UTC, datetime
 from pathlib import Path
+from typing import Any
 
-from sqlalchemy import func, select
+from sqlalchemy import func, select, update
 from sqlalchemy.orm import Session
 
-from ..models import Artist, ImageRecord, Tag
+from ..models import Artist, ImageRecord, LibraryAuditRun, Tag
 from ..models.series_page import SeriesPage
 from ..models.tag import image_tag
 
@@ -365,3 +367,146 @@ def prune_unused_tags(session: Session, *, dry_run: bool = False) -> dict:
         )
         session.commit()
     return {"deleted": len(ids), "sample_names": sample}
+
+
+# ---------------------------------------------------------------------------
+# FC-Cleanup additions (2026-05-26): retroactive audit of import-filter rules.
+# ---------------------------------------------------------------------------
+
+_MIN_DIM_SAMPLE_CAP = 50
+
+
+def project_min_dimension_violations(
+    session: Session, *, min_width: int, min_height: int,
+) -> dict:
+    """Return {count, sample_ids} for image_record rows with width or
+    height below the thresholds. Synchronous SQL — no PIL inspection
+    needed since width/height are stored columns."""
+    base = select(ImageRecord.id).where(
+        (ImageRecord.width < min_width) | (ImageRecord.height < min_height)
+    )
+    count = session.execute(
+        select(func.count()).select_from(base.subquery())
+    ).scalar_one()
+    sample_ids = session.execute(
+        base.order_by(ImageRecord.id).limit(_MIN_DIM_SAMPLE_CAP)
+    ).scalars().all()
+    return {"count": count, "sample_ids": list(sample_ids)}
+
+
+def delete_min_dimension_violations(
+    session: Session, *, min_width: int, min_height: int, images_root: Path,
+) -> int:
+    """Delete every image_record where width<min_w OR height<min_h.
+    Routes through delete_images so file-unlink + cascading FKs
+    (image_tag / image_provenance / etc.) are handled uniformly."""
+    ids = session.execute(
+        select(ImageRecord.id).where(
+            (ImageRecord.width < min_width) | (ImageRecord.height < min_height)
+        )
+    ).scalars().all()
+    if not ids:
+        return 0
+    result = delete_images(
+        session, image_ids=list(ids), images_root=images_root,
+    )
+    return result["images_deleted"]
+
+
+# ---------------------------------------------------------------------------
+# Audit lifecycle (transparency + single_color async scans).
+# ---------------------------------------------------------------------------
+
+
+class AuditAlreadyRunning(Exception):
+    """Another audit_run is currently in status='running' — wait or
+    cancel it before starting a new one. Surfaces as HTTP 409 in the
+    /api/cleanup/audit POST endpoint."""
+
+
+class AuditNotReady(Exception):
+    """apply_audit_run called on an audit whose status is not 'ready'."""
+
+
+class ConfirmTokenMismatch(Exception):
+    """Operator-supplied confirm token did not match server-recomputed token."""
+
+
+_VALID_RULES = ("transparency", "single_color")
+
+
+def start_audit_run(
+    session: Session, *, rule: str, params: dict[str, Any],
+) -> int:
+    """Create a LibraryAuditRun row in status='running' and dispatch the
+    scan_library_for_rule Celery task. Returns the new audit_id.
+
+    Concurrent-runs guard: raises AuditAlreadyRunning if any audit_run
+    has status='running'. Operator must cancel or wait."""
+    if rule not in _VALID_RULES:
+        raise ValueError(f"unknown rule {rule!r}; expected one of {_VALID_RULES}")
+    existing = session.execute(
+        select(LibraryAuditRun.id).where(LibraryAuditRun.status == "running")
+    ).scalar_one_or_none()
+    if existing is not None:
+        raise AuditAlreadyRunning(existing)
+    audit = LibraryAuditRun(
+        rule=rule,
+        params=params,
+        status="running",
+        scanned_count=0,
+        matched_count=0,
+        matched_ids=[],
+    )
+    session.add(audit)
+    session.flush()
+    audit_id = audit.id
+    # Dispatch after flush so audit_id is populated; commit happens in
+    # the API handler so the audit row + dispatch are visible together.
+    from ..tasks.library_audit import scan_library_for_rule
+    scan_library_for_rule.delay(audit_id)
+    return audit_id
+
+
+def apply_audit_run(
+    session: Session, *, audit_id: int, confirm_token: str, images_root: Path,
+) -> int:
+    """Delete all images in audit_run.matched_ids after confirming token.
+    Marks audit status='applied'. Routes through delete_images so files
+    + cascading FK rows are handled uniformly."""
+    audit = session.execute(
+        select(LibraryAuditRun).where(LibraryAuditRun.id == audit_id)
+    ).scalar_one_or_none()
+    if audit is None:
+        raise ValueError(f"audit_run {audit_id} not found")
+    if audit.status != "ready":
+        raise AuditNotReady(audit.status)
+    # Token format matches modal/DestructiveConfirmModal.vue convention:
+    # ${action}-${kind}-${runId}. The modal hardcodes action ∈ {'restore',
+    # 'delete'}; "apply audit" is semantically a delete of the matched
+    # images, so we use 'delete-audit-<id>' (not 'apply-audit-<id>').
+    expected = f"delete-audit-{audit_id}"
+    if confirm_token != expected:
+        raise ConfirmTokenMismatch(expected)
+    ids = list(audit.matched_ids or [])
+    deleted = 0
+    if ids:
+        result = delete_images(session, image_ids=ids, images_root=images_root)
+        deleted = result["images_deleted"]
+    session.execute(
+        update(LibraryAuditRun)
+        .where(LibraryAuditRun.id == audit_id)
+        .values(status="applied", finished_at=datetime.now(UTC))
+    )
+    return deleted
+
+
+def cancel_audit_run(session: Session, *, audit_id: int) -> None:
+    """Flip a running audit_run to 'cancelled'. The scan task checks
+    for status=='cancelled' between batches and exits cleanly."""
+    session.execute(
+        update(LibraryAuditRun)
+        .where(LibraryAuditRun.id == audit_id)
+        .where(LibraryAuditRun.status == "running")
+        .values(status="cancelled", finished_at=datetime.now(UTC))
+    )

backend/app/services/importer.py

@@ -18,6 +18,7 @@ from pathlib import Path
 
 from PIL import Image
 from sqlalchemy import select
+from sqlalchemy.exc import IntegrityError
 from sqlalchemy.orm import Session
 
 from ..models import (
@@ -51,7 +52,14 @@ class SkipReason(StrEnum):
 
 @dataclass(frozen=True)
 class ImportResult:
-    status: str  # 'imported'|'skipped'|'failed'|'superseded'|'attached'
+    # 'imported'   — new ImageRecord row created
+    # 'superseded' — existing ImageRecord row got the new file (larger) + sidecar
+    # 'attached'   — non-media saved as PostAttachment
+    # 'refreshed'  — deep scan re-applied sidecar / filled NULL phash / NULL
+    #                artist on an already-imported row (no new ImageRecord)
+    # 'skipped'    — no work done (true duplicate, too small, etc.)
+    # 'failed'     — pipeline error
+    status: str
     image_id: int | None = None
     skip_reason: SkipReason | None = None
     error: str | None = None
@@ -71,6 +79,31 @@ def is_video(path: Path) -> bool:
     return path.suffix.lower() in VIDEO_EXTS
 
 
+def _safe_ext(path: Path) -> str:
+    """Conservatively extract a file extension for PostAttachment.ext
+    (varchar(32)).
+
+    gallery-dl produces some filenames with URL-encoded query-string
+    artifacts embedded into the basename (e.g.
+    `79507046_media_..._https___www.patreon.com_media-u_Z0FBQUFBQm5q...`).
+    `Path.suffix` finds the LAST dot and returns everything after, which
+    in those cases yields a 50+ char "extension" of mostly base64-ish
+    junk. That blows the column. Operator-flagged 2026-05-25.
+
+    Real extensions are short and alphanumeric. We accept anything ≤ 16
+    chars where every post-dot character is alphanumeric; anything else
+    means the input wasn't a real extension and we return the empty
+    string. ext is nullable-ish (empty string still satisfies NOT NULL)
+    and consumers should treat "" as "no known extension".
+    """
+    suffix = path.suffix.lower()
+    if not suffix or len(suffix) > 16:
+        return ""
+    if not all(c.isalnum() for c in suffix[1:]):
+        return ""
+    return suffix
+
+
 def _mime_for(path: Path) -> str:
     suffix = path.suffix.lower()
     image_mimes = {
@@ -126,6 +159,123 @@ class Importer:
         self.settings = settings
         self.deep = deep
         self.attachments = AttachmentStore(images_root)
+        # phash near-dup candidate cache. Archive imports call _import_media
+        # per-member; without this cache the per-member SELECT *FROM
+        # image_record WHERE phash IS NOT NULL fetch repeats N times and a
+        # large library × many-member archive blew past soft_time_limit
+        # (300s) — operator-flagged 2026-05-25. Loaded lazily on first
+        # need, appended to on every imported/superseded outcome, never
+        # invalidated mid-Importer (Importer instances are per-task /
+        # per-archive-import so cross-instance staleness is harmless).
+        self._phash_candidates: list[tuple] | None = None
+
+    def _phash_candidates_cache(self) -> list[tuple]:
+        """Cached `(phash, width, height, id)` rows from image_record.
+        Loaded on first call, appended-to on subsequent imported/
+        superseded outcomes. Soft-timeout pattern: an archive with N
+        members + a library of M existing rows used to do N × M-row
+        fetches (operator-flagged 2026-05-25); now it's exactly one.
+
+        The per-task lifecycle of Importer (instantiated fresh by
+        import_media_file) bounds the cache's staleness window: cross-
+        process changes (other workers importing concurrently) won't
+        be reflected, but that's the same race the un-cached version
+        had — `find_similar` is best-effort anyway."""
+        if self._phash_candidates is None:
+            rows = self.session.execute(
+                select(
+                    ImageRecord.phash,
+                    ImageRecord.width,
+                    ImageRecord.height,
+                    ImageRecord.id,
+                ).where(ImageRecord.phash.is_not(None))
+            ).all()
+            self._phash_candidates = [
+                (r.phash, r.width or 0, r.height or 0, r.id) for r in rows
+            ]
+        return self._phash_candidates
+
+    def _phash_cache_append(self, phash, width, height, image_id) -> None:
+        """Append a freshly-imported row to the cache so subsequent
+        members of the same archive can match against it."""
+        if self._phash_candidates is not None and phash is not None:
+            self._phash_candidates.append(
+                (phash, width or 0, height or 0, image_id)
+            )
+
+    def _find_or_create_source(
+        self, *, artist_id: int, platform: str, url: str,
+    ) -> Source:
+        """Race-safe find-or-create on `source` keyed by
+        (artist_id, platform, url) — the same key as the
+        `uq_source_artist_platform_url` constraint.
+
+        Two concurrent workers processing different files in the same
+        post can both find no existing Source row then both INSERT,
+        which trips the unique constraint and poisons the session with
+        `psycopg.errors.UniqueViolation`. Operator-flagged 2026-05-26.
+
+        Pattern: select; if absent, open a savepoint and INSERT.
+        On IntegrityError, roll the savepoint back (NOT the outer
+        transaction, which would lose the surrounding scan's progress)
+        and re-select — the concurrent op just created the row we
+        wanted, so the second select will find it.
+        """
+        existing = self.session.execute(
+            select(Source).where(
+                Source.artist_id == artist_id,
+                Source.platform == platform,
+                Source.url == url,
+            )
+        ).scalar_one_or_none()
+        if existing is not None:
+            return existing
+        sp = self.session.begin_nested()
+        try:
+            row = Source(artist_id=artist_id, platform=platform, url=url)
+            self.session.add(row)
+            self.session.flush()
+            sp.commit()
+            return row
+        except IntegrityError:
+            sp.rollback()
+            return self.session.execute(
+                select(Source).where(
+                    Source.artist_id == artist_id,
+                    Source.platform == platform,
+                    Source.url == url,
+                )
+            ).scalar_one()
+
+    def _find_or_create_post(
+        self, *, source_id: int, external_post_id: str,
+    ) -> Post:
+        """Race-safe find-or-create on `post` keyed by
+        (source_id, external_post_id). Mirrors `_find_or_create_source`
+        — same savepoint + IntegrityError-recovery pattern."""
+        existing = self.session.execute(
+            select(Post).where(
+                Post.source_id == source_id,
+                Post.external_post_id == external_post_id,
+            )
+        ).scalar_one_or_none()
+        if existing is not None:
+            return existing
+        sp = self.session.begin_nested()
+        try:
+            row = Post(source_id=source_id, external_post_id=external_post_id)
+            self.session.add(row)
+            self.session.flush()
+            sp.commit()
+            return row
+        except IntegrityError:
+            sp.rollback()
+            return self.session.execute(
+                select(Post).where(
+                    Post.source_id == source_id,
+                    Post.external_post_id == external_post_id,
+                )
+            ).scalar_one()
 
     def import_one(self, source: Path) -> ImportResult:
         """Dispatch by kind. Media → normal pipeline. Archive → extract
@@ -166,29 +316,13 @@ class Importer:
         sd = parse_sidecar(data)
         platform = sd.platform or "unknown"
         url = sd.post_url or f"sidecar:{platform}"
-        src = self.session.execute(
-            select(Source).where(
-                Source.artist_id == artist.id,
-                Source.platform == platform,
-                Source.url == url,
-            )
-        ).scalar_one_or_none()
-        if src is None:
-            src = Source(artist_id=artist.id, platform=platform, url=url)
-            self.session.add(src)
-            self.session.flush()
+        src = self._find_or_create_source(
+            artist_id=artist.id, platform=platform, url=url,
+        )
         epid = sd.external_post_id or sc.stem
-        post = self.session.execute(
-            select(Post).where(
-                Post.source_id == src.id,
-                Post.external_post_id == epid,
-            )
-        ).scalar_one_or_none()
-        if post is None:
-            post = Post(source_id=src.id, external_post_id=epid)
-            self.session.add(post)
-            self.session.flush()
-        return post
+        return self._find_or_create_post(
+            source_id=src.id, external_post_id=epid,
+        )
 
     def _capture_attachment(
         self, source: Path, *, post: Post | None = None,
@@ -209,7 +343,7 @@ class Importer:
                 sha256=sha,
                 path=stored,
                 original_filename=source.name,
-                ext=source.suffix.lower(),
+                ext=_safe_ext(source),
                 mime=_mime_for(source),
                 size_bytes=source.stat().st_size,
             ))
@@ -324,18 +458,7 @@ class Importer:
                     error=f"PIL load failed during phash compute: {exc}",
                 )
             if phash is not None:
-                cand_rows = self.session.execute(
-                    select(
-                        ImageRecord.phash,
-                        ImageRecord.width,
-                        ImageRecord.height,
-                        ImageRecord.id,
-                    ).where(ImageRecord.phash.is_not(None))
-                ).all()
-                candidates = [
-                    (c.phash, c.width or 0, c.height or 0, c.id)
-                    for c in cand_rows
-                ]
+                candidates = self._phash_candidates_cache()
                 rel, match_id = find_similar(
                     phash, width or 0, height or 0,
                     candidates, self.settings.phash_threshold,
@@ -369,6 +492,7 @@ class Importer:
         )
         self.session.add(record)
         self.session.flush()
+        self._phash_cache_append(phash, width, height, record.id)
 
         # Folder→artist (anchored to attribution_path).
         artist = None
@@ -392,13 +516,27 @@ class Importer:
     ) -> ImportResult:
         """Deep scan: backfill phash/provenance/artist on an
         already-imported record. METADATA ONLY — never re-runs the pHash
-        near-dup / supersede path. NULL-only, idempotent."""
+        near-dup / supersede path. NULL-only on phash/artist, additive on
+        sidecar Post/Source/ImageProvenance (via _apply_sidecar).
+        Idempotent: a second deep-scan over the same file finds nothing
+        to refresh and is a no-op.
+
+        Returns status="refreshed" so the UI can surface the work done
+        instead of the prior misleading "skipped/duplicate_hash" reading.
+        Operator-flagged 2026-05-25 — IR has had this; FC inherited it
+        as a no-op skip during the original port and the UI showed deep
+        scan as "completed with no changes" even when sidecar metadata
+        actually got re-applied to N existing rows.
+        """
         if existing.phash is None and not is_video(source):
             try:
                 with Image.open(source) as im:
                     ph = compute_phash(im)
                 if ph is not None:
                     existing.phash = ph
+                    # Promoted from NULL to non-NULL → cache is now stale
+                    # (this row would newly qualify for the candidates set).
+                    self._phash_candidates = None
             except Exception as exc:
                 log.warning("deep rephash failed for %s: %s", source, exc)
 
@@ -411,10 +549,7 @@ class Importer:
 
         self._apply_sidecar(existing, attribution_path, artist)
         self.session.commit()
-        return ImportResult(
-            status="skipped", skip_reason=SkipReason.duplicate_hash,
-            image_id=existing.id, error="deep: re-derived",
-        )
+        return ImportResult(status="refreshed", image_id=existing.id)
 
     def attach_in_place(
         self,
@@ -498,18 +633,7 @@ class Importer:
             except Exception:
                 phash = None
             if phash is not None:
-                cand_rows = self.session.execute(
-                    select(
-                        ImageRecord.phash,
-                        ImageRecord.width,
-                        ImageRecord.height,
-                        ImageRecord.id,
-                    ).where(ImageRecord.phash.is_not(None))
-                ).all()
-                candidates = [
-                    (c.phash, c.width or 0, c.height or 0, c.id)
-                    for c in cand_rows
-                ]
+                candidates = self._phash_candidates_cache()
                 rel, match_id = find_similar(
                     phash, width or 0, height or 0,
                     candidates, self.settings.phash_threshold,
@@ -544,6 +668,7 @@ class Importer:
             record.artist_id = artist.id
         self.session.add(record)
         self.session.flush()
+        self._phash_cache_append(phash, width, height, record.id)
 
         # Sidecar provenance (best-effort). When `source` is passed, link
         # the post to that subscription Source instead of creating a new
@@ -639,29 +764,14 @@ class Importer:
         else:
             platform = sd.platform or "unknown"
             url = sd.post_url or f"sidecar:{platform}"
-            src = self.session.execute(
-                select(Source).where(
-                    Source.artist_id == artist.id,
-                    Source.platform == platform,
-                    Source.url == url,
-                )
-            ).scalar_one_or_none()
-            if src is None:
-                src = Source(artist_id=artist.id, platform=platform, url=url)
-                self.session.add(src)
-                self.session.flush()
+            src = self._find_or_create_source(
+                artist_id=artist.id, platform=platform, url=url,
+            )
 
         epid = sd.external_post_id or sc.stem
-        post = self.session.execute(
-            select(Post).where(
-                Post.source_id == src.id,
-                Post.external_post_id == epid,
-            )
-        ).scalar_one_or_none()
-        if post is None:
-            post = Post(source_id=src.id, external_post_id=epid)
-            self.session.add(post)
-            self.session.flush()
+        post = self._find_or_create_post(
+            source_id=src.id, external_post_id=epid,
+        )
         if sd.post_url is not None:
             post.post_url = sd.post_url
         if sd.post_title is not None:
@@ -722,6 +832,15 @@ class Importer:
         row id (so tags/series/curation stay attached). ML is cleared so
         the import task re-derives it on the new pixels.
 
+        After the file swap, the new file's adjacent gallery-dl sidecar
+        (if any) is applied via _apply_sidecar — operator-flagged
+        2026-05-25: scanning a GS download dir with smaller IR-migrated
+        images on the receiving end used to swap files but lose the GS
+        sidecar's post metadata entirely. _apply_sidecar is additive
+        (find-or-create Post / Source / ImageProvenance, NULL-only
+        primary_post_id update) so any pre-existing Post linkage
+        survives untouched.
+
         If `new_path` is provided, `source` is assumed to ALREADY be at
         that path (FC-3c attach_in_place case) — skip the copy step.
         Otherwise the file is copied via _copy_to_library."""
@@ -750,6 +869,26 @@ class Importer:
         # created_at intentionally preserved; updated_at auto-bumps.
         self.session.flush()
         self.session.commit()
+        # The phash candidate cache (used to avoid N+1 selects during
+        # archive imports) is now stale for `existing.id` — the row's
+        # phash/dimensions changed. Invalidate; the next call re-fetches.
+        self._phash_candidates = None
+
+        # Sidecar enrichment from the new (larger) file's location.
+        # _apply_sidecar resolves artist from the sidecar itself if the
+        # existing row has none, and is internally guarded against
+        # missing-or-malformed sidecars (silent return).
+        try:
+            self._apply_sidecar(existing, source, None)
+        except Exception as exc:
+            # Don't unwind the supersede DB swap if sidecar parsing
+            # blows up unexpectedly — the file replacement is the
+            # critical operation, sidecar is enrichment.
+            log.warning(
+                "sidecar enrichment failed during supersede of "
+                "image_record.id=%s from %s: %s",
+                existing.id, source, exc,
+            )
 
         for stale in (old_path, old_thumb):
             if not stale or stale == str(dest):
@@ -765,8 +904,26 @@ class Importer:
                 pass
 
     def _transparency_pct(self, source: Path) -> float:
-        """Fraction of fully-transparent pixels in the image. 0.0 if no alpha."""
+        """Fraction of fully-transparent pixels in the image. 0.0 if no alpha.
+
+        For animated formats (multi-frame WebP / GIF / APNG), short-circuit
+        to 0.0 instead of decoding every frame. PIL's `getchannel("A")`
+        forces a full decode of all frames in an animated image, which for
+        a large animated WebP takes 5+ minutes and blows past the Celery
+        soft+hard time limits (300s/360s → SIGKILL). Operator-flagged
+        2026-05-26. Transparency analysis on a multi-frame image isn't
+        meaningful for art-curation purposes anyway — different frames
+        have different alpha — so the existing too_transparent skip rule
+        is bypassed entirely for animated content.
+        """
         with Image.open(source) as im:
+            if getattr(im, "is_animated", False):
+                log.info(
+                    "skipping transparency check for animated image %s "
+                    "(n_frames=%d) — avoids multi-frame decode timeout",
+                    source, getattr(im, "n_frames", 0),
+                )
+                return 0.0
             if im.mode not in ("RGBA", "LA") and not (
                 im.mode == "P" and "transparency" in im.info
             ):

backend/app/services/ml/embedder.py

@@ -34,10 +34,21 @@ class Embedder:
         if self._model is not None:
             return
         import torch
-        from transformers import AutoModel, AutoProcessor
+        from transformers import AutoModel, SiglipImageProcessor
 
         self._torch = torch
-        self._processor = AutoProcessor.from_pretrained(str(self._model_dir))
+        # FC's embedder only does IMAGE inference — never text. AutoProcessor
+        # loads the full processor including SiglipTokenizer, which requires
+        # the sentencepiece library at import time even if we never call it.
+        # SiglipImageProcessor loads ONLY preprocessor_config.json (image
+        # side) and skips the tokenizer config entirely. Operator hit the
+        # ImportError 2026-05-25 once the ml-worker started actually running
+        # tag_and_embed; switching to the image-only loader avoids the
+        # tokenizer dep without adding ~30 MB of unused C++ build to the
+        # lean ml-worker image.
+        self._processor = SiglipImageProcessor.from_pretrained(
+            str(self._model_dir)
+        )
         self._model = AutoModel.from_pretrained(str(self._model_dir))
         self._model.eval()
 

backend/app/tasks/import_file.py

@@ -29,10 +29,12 @@ IMAGES_ROOT = Path("/images")
 def _map_result_to_status(result):
     """(ImportTask.status, should_requeue_ml_and_thumb) for an ImportResult.
     'superseded' = the kept row's file/ML changed → complete + re-derive.
-    'attached' = a non-art file preserved → complete, no ML/thumb."""
+    'attached' = a non-art file preserved → complete, no ML/thumb.
+    'refreshed' = deep scan refreshed sidecar/phash on an existing row →
+    complete, no ML/thumb re-derive (file/pixels unchanged)."""
     if result.status in ("imported", "superseded"):
         return ("complete", True)
-    if result.status == "attached":
+    if result.status in ("attached", "refreshed"):
         return ("complete", False)
     if result.status == "skipped":
         return ("skipped", False)
@@ -138,6 +140,15 @@ def _do_import(session, task, import_task_id: int) -> dict:
         task.result_image_id = result.image_id
         counter_col_name = "imported"
         counter_col = ImportBatch.imported
+    elif result.status == "refreshed":
+        # Deep-scan rederive: existing row got phash/artist/sidecar
+        # refreshed. Task is complete (no further work), but counted in
+        # `refreshed` not `imported` so the UI can surface the actual
+        # work done. operator-flagged 2026-05-25.
+        task.status = "complete"
+        task.result_image_id = result.image_id
+        counter_col_name = "refreshed"
+        counter_col = ImportBatch.refreshed
     elif result.status == "attached":
         task.status = "complete"
         counter_col_name = "attachments"

backend/app/tasks/library_audit.py

@@ -0,0 +1,167 @@
+"""scan_library_for_rule Celery task — iterates image_record in keyset-
+paginated batches, evaluates the audit rule per image, populates
+LibraryAuditRun.matched_ids. Runs on the maintenance queue with a 2h soft
+time limit (plenty of margin for 100k+ image libraries at ~100ms PIL
+decode + histogram per image).
+
+State machine:
+  start: status='running'
+  end success: status='ready'
+  end error: status='error', error=traceback
+  oversize: status='error', error='matched too many images; tighten threshold'
+  external cancel: scan sees status='cancelled' between batches, exits.
+"""
+
+import logging
+import traceback
+from datetime import UTC, datetime
+
+from celery.exceptions import SoftTimeLimitExceeded
+from PIL import Image
+from sqlalchemy import select, update
+from sqlalchemy.exc import DBAPIError, OperationalError
+
+from ..celery_app import celery
+from ..models import ImageRecord, LibraryAuditRun
+from ..services.audits import single_color, transparency
+from ._sync_engine import sync_session_factory as _sync_session_factory
+
+log = logging.getLogger(__name__)
+
+_BATCH = 500
+_PROGRESS_TICK = 100
+_MAX_MATCHED = 50_000
+
+_RULES = {
+    "transparency": transparency.evaluate,
+    "single_color": single_color.evaluate,
+}
+
+
+@celery.task(
+    name="backend.app.tasks.library_audit.scan_library_for_rule",
+    bind=True,
+    autoretry_for=(OperationalError, DBAPIError),
+    retry_backoff=5,
+    retry_backoff_max=60,
+    retry_jitter=True,
+    max_retries=3,
+    soft_time_limit=7200,
+    time_limit=7500,
+)
+def scan_library_for_rule(self, audit_id: int) -> dict:
+    """See module docstring. Returns a small summary dict for eager-mode
+    test assertions (real workers ignore the return value)."""
+    SessionLocal = _sync_session_factory()
+    try:
+        with SessionLocal() as session:
+            audit = session.get(LibraryAuditRun, audit_id)
+            if audit is None:
+                return {"audit_id": audit_id, "status": "missing"}
+            evaluate = _RULES.get(audit.rule)
+            if evaluate is None:
+                _mark_error(session, audit_id, f"unknown rule {audit.rule!r}")
+                return {"audit_id": audit_id, "status": "error"}
+            params = dict(audit.params or {})
+            matched: list[int] = []
+            scanned = 0
+            last_id = 0
+            while True:
+                # Cancellation check between batches.
+                current_status = session.execute(
+                    select(LibraryAuditRun.status)
+                    .where(LibraryAuditRun.id == audit_id)
+                ).scalar_one()
+                if current_status == "cancelled":
+                    return {"audit_id": audit_id, "status": "cancelled"}
+                rows = session.execute(
+                    select(ImageRecord.id, ImageRecord.path)
+                    .where(ImageRecord.id > last_id)
+                    .where(ImageRecord.mime.like("image/%"))
+                    .order_by(ImageRecord.id.asc())
+                    .limit(_BATCH)
+                ).all()
+                if not rows:
+                    break
+                for image_id, image_path in rows:
+                    last_id = image_id
+                    scanned += 1
+                    try:
+                        with Image.open(image_path) as im:
+                            try:
+                                if evaluate(im, **params):
+                                    matched.append(image_id)
+                            except Exception as exc:  # noqa: BLE001
+                                log.warning(
+                                    "audit %s: rule evaluate failed on %s: %s",
+                                    audit_id, image_path, exc,
+                                )
+                    except FileNotFoundError:
+                        log.warning(
+                            "audit %s: image_record %s file missing at %s; skipping",
+                            audit_id, image_id, image_path,
+                        )
+                    except OSError as exc:
+                        log.warning(
+                            "audit %s: PIL load failed for %s: %s",
+                            audit_id, image_path, exc,
+                        )
+                    if len(matched) > _MAX_MATCHED:
+                        _mark_error(
+                            session, audit_id,
+                            f"matched > {_MAX_MATCHED} images; "
+                            "tighten threshold and re-run",
+                        )
+                        return {"audit_id": audit_id, "status": "error"}
+                    if scanned % _PROGRESS_TICK == 0:
+                        session.execute(
+                            update(LibraryAuditRun)
+                            .where(LibraryAuditRun.id == audit_id)
+                            .values(scanned_count=scanned)
+                        )
+                        session.commit()
+            # Final state.
+            session.execute(
+                update(LibraryAuditRun)
+                .where(LibraryAuditRun.id == audit_id)
+                .values(
+                    scanned_count=scanned,
+                    matched_count=len(matched),
+                    matched_ids=matched,
+                    status="ready",
+                    finished_at=datetime.now(UTC),
+                )
+            )
+            session.commit()
+            return {
+                "audit_id": audit_id,
+                "status": "ready",
+                "scanned": scanned,
+                "matched": len(matched),
+            }
+    except SoftTimeLimitExceeded:
+        with SessionLocal() as session:
+            _mark_error(session, audit_id, "soft_time_limit exceeded (>7200s)")
+        raise
+    except (OperationalError, DBAPIError):
+        # Retryable per the decorator; leave row in 'running' and let
+        # autoretry try again. Recovery sweep catches if all retries fail.
+        raise
+    except Exception:  # noqa: BLE001
+        tb = traceback.format_exc()
+        with SessionLocal() as session:
+            _mark_error(session, audit_id, tb)
+        raise
+
+
+def _mark_error(session, audit_id: int, error_msg: str) -> None:
+    session.execute(
+        update(LibraryAuditRun)
+        .where(LibraryAuditRun.id == audit_id)
+        .values(
+            status="error",
+            error=error_msg,
+            finished_at=datetime.now(UTC),
+        )
+    )
+    session.commit()

backend/app/tasks/scan.py

@@ -59,16 +59,25 @@ def scan_directory(self, triggered_by: str = "manual",
         session.flush()
         batch_id = batch.id
 
-        # Skip-set: any source_path that already has a non-failed ImportTask
-        # row. Re-running scan_directory must not re-enqueue files the
-        # importer has already handled (or is currently handling); doing so
-        # creates duplicate work and inflates the queue. Failed prior tasks
-        # are eligible for retry.
+        # Skip-set behavior splits by mode (operator-flagged 2026-05-25):
+        #
+        # quick: any non-failed prior ImportTask (active OR finished) is
+        #   skipped — quick scan only does new-file enqueue, so re-touching
+        #   already-imported files is wasted work.
+        #
+        # deep: ONLY currently-in-flight tasks (pending/queued/processing)
+        #   are skipped. Completed and skipped tasks ARE re-queued because
+        #   deep scan exists precisely to re-touch already-imported files
+        #   (refresh sidecar metadata, fill NULL phash, fill NULL artist
+        #   via Importer._deep_rederive). Matches IR's deep-scan behavior.
+        active_statuses = ["pending", "queued", "processing"]
+        if mode == "deep":
+            skip_statuses = active_statuses
+        else:
+            skip_statuses = active_statuses + ["complete", "skipped"]
         non_failed_existing = set(session.execute(
             select(ImportTask.source_path).where(
-                ImportTask.status.in_(
-                    ["pending", "queued", "processing", "complete", "skipped"]
-                ),
+                ImportTask.status.in_(skip_statuses),
             )
         ).scalars().all())
 

backend/app/utils/phash.py

@@ -13,8 +13,21 @@ HASH_SIZE = 8
 
 def compute_phash(pil_image) -> str | None:
     """Perceptual hash of an opened PIL image, as a hex string. None on any
-    failure (videos/unreadable/non-image)."""
+    failure (videos/unreadable/non-image).
+
+    For animated images (multi-frame WebP/GIF/APNG), explicitly seek to
+    frame 0 first. Without this, some PIL operations downstream of
+    imagehash.phash (convert("L"), resize) can iterate all frames and
+    blow past Celery's hard time limit on large animations
+    (operator-flagged 2026-05-26 against animated WebPs). The pHash of
+    frame 0 is the conventional choice for animated content.
+    """
     try:
+        if getattr(pil_image, "is_animated", False):
+            try:
+                pil_image.seek(0)
+            except Exception:
+                pass
         return str(imagehash.phash(pil_image, hash_size=HASH_SIZE))
     except Exception:
         return None

backend/app/utils/sidecar.py

@@ -4,6 +4,7 @@ No per-platform branching: a small common key set with fallbacks; the
 full JSON is kept in raw so anything unmapped is recoverable later.
 """
 
+import re
 from dataclasses import dataclass
 from datetime import UTC, datetime
 from pathlib import Path
@@ -21,10 +22,28 @@ class SidecarData:
     raw: dict
 
 
+# gallery-dl prefixes media filenames with `NN_` for in-post ordering
+# (`01_HOLLOW-ICHIGO.png`, `02_HOLOW ICHIGO.zip`) but writes the sidecar
+# under the attachment's stem WITHOUT that ordering prefix
+# (`HOLLOW-ICHIGO.json`). Strip the prefix when looking for sidecars.
+# Confirmed against real Patreon downloads 2026-05-26 — without this
+# strip, every gallery-dl post-level sidecar was invisible to FC since
+# FC-3 shipped (24 deep-refresh calls produced 0 Posts in operator's DB).
+_NUMBERING_PREFIX = re.compile(r"^\d+_(.+)$")
+
+
 def find_sidecar(media: Path) -> Path | None:
+    # Attachment-level sidecars (image.jpg.json, image.json).
     for cand in (media.with_suffix(".json"), Path(str(media) + ".json")):
         if cand.is_file():
             return cand
+    # gallery-dl post-numbered convention: strip the `NN_` prefix from
+    # the stem and look for that.json in the same directory.
+    m = _NUMBERING_PREFIX.match(media.stem)
+    if m:
+        cand = media.parent / f"{m.group(1)}.json"
+        if cand.is_file():
+            return cand
     return None
 
 

extension/manifest.json

@@ -1,7 +1,7 @@
 {
   "manifest_version": 3,
   "name": "FabledCurator",
-  "version": "1.0.0",
+  "version": "1.0.3",
   "description": "Export cookies from supported platforms to FabledCurator and add creators as sources in one click.",
 
   "browser_specific_settings": {

extension/package.json

@@ -1,13 +1,13 @@
 {
   "name": "fabledcurator-extension",
-  "version": "1.0.0",
+  "version": "1.0.3",
   "private": true,
   "description": "Firefox extension for FabledCurator",
   "scripts": {
-    "lint":  "web-ext lint --source-dir=.",
-    "start": "web-ext run --source-dir=. --firefox=firefox",
-    "build": "web-ext build --source-dir=. --overwrite-dest",
-    "sign":  "web-ext sign --source-dir=. --channel=unlisted --api-key=$WEB_EXT_API_KEY --api-secret=$WEB_EXT_API_SECRET"
+    "lint":  "web-ext lint --source-dir=. --no-config-discovery --ignore-files package.json package-lock.json web-ext-artifacts node_modules README.md .gitignore",
+    "start": "web-ext run --source-dir=. --no-config-discovery --ignore-files package.json package-lock.json web-ext-artifacts node_modules README.md .gitignore --firefox=firefox",
+    "build": "web-ext build --source-dir=. --no-config-discovery --ignore-files package.json package-lock.json web-ext-artifacts node_modules README.md .gitignore --overwrite-dest",
+    "sign":  "web-ext sign --source-dir=. --no-config-discovery --ignore-files package.json package-lock.json web-ext-artifacts node_modules README.md .gitignore --channel=unlisted --api-key=$WEB_EXT_API_KEY --api-secret=$WEB_EXT_API_SECRET"
   },
   "devDependencies": {
     "web-ext": "^8.0.0"

extension/web-ext-config.cjs

@@ -1,13 +0,0 @@
-module.exports = {
-  sourceDir: '.',
-  artifactsDir: './web-ext-artifacts',
-  ignoreFiles: [
-    'package.json',
-    'package-lock.json',
-    'web-ext-config.cjs',
-    'web-ext-artifacts',
-    'node_modules',
-    'README.md',
-    '.gitignore',
-  ],
-};

frontend/src/components/cleanup/MinDimensionCard.vue

@@ -0,0 +1,127 @@
+<template>
+  <v-card class="fc-clean-card">
+    <v-card-title class="d-flex align-center" style="gap: 10px;">
+      <v-icon icon="mdi-image-size-select-small" size="small" />
+      <span>Minimum dimensions</span>
+    </v-card-title>
+    <v-card-text>
+      <p class="fc-muted text-body-2 mb-3">
+        Find and delete images smaller than the threshold. Mirrors the
+        import-time <code>min_width</code> / <code>min_height</code>
+        filter, applied retroactively to the existing library.
+      </p>
+
+      <v-row dense>
+        <v-col cols="6">
+          <v-text-field
+            v-model.number="minW" label="Min width (px)" type="number"
+            min="0" density="compact" hide-details
+          />
+        </v-col>
+        <v-col cols="6">
+          <v-text-field
+            v-model.number="minH" label="Min height (px)" type="number"
+            min="0" density="compact" hide-details
+          />
+        </v-col>
+      </v-row>
+
+      <div class="d-flex align-center mt-3" style="gap: 10px;">
+        <v-btn
+          color="accent" variant="flat" rounded="pill"
+          prepend-icon="mdi-magnify"
+          :loading="busy"
+          @click="onPreview"
+        >Preview</v-btn>
+
+        <span v-if="preview" class="text-body-2">
+          <strong>{{ preview.count }}</strong> image(s) would be deleted.
+        </span>
+      </div>
+
+      <v-btn
+        v-if="preview && preview.count > 0"
+        class="mt-3"
+        color="error" variant="flat" rounded="pill"
+        prepend-icon="mdi-delete"
+        @click="onDeleteClick"
+      >Delete {{ preview.count }} matching...</v-btn>
+    </v-card-text>
+
+    <DestructiveConfirmModal
+      v-model="showModal"
+      action="delete"
+      kind="min-dim"
+      :run-id="tokenSha8"
+      tier="C"
+      :projected-counts="projectedCounts"
+      :description="`Width < ${minW} OR height < ${minH}`"
+      @confirm="onConfirmedDelete"
+    />
+  </v-card>
+</template>
+
+<script setup>
+import { onMounted, ref } from 'vue'
+
+import DestructiveConfirmModal from '../modal/DestructiveConfirmModal.vue'
+import { useCleanupStore } from '../../stores/cleanup.js'
+
+const store = useCleanupStore()
+const minW = ref(0)
+const minH = ref(0)
+const preview = ref(null)
+const busy = ref(false)
+const showModal = ref(false)
+const tokenSha8 = ref('')
+const projectedCounts = ref({})
+
+onMounted(async () => {
+  await store.loadDefaults()
+  minW.value = store.defaults.min_width
+  minH.value = store.defaults.min_height
+})
+
+// SHA-256 truncated to 8 hex chars — matches the backend's
+// _min_dim_token() exactly. Web Crypto rejects MD5 as insecure.
+async function sha8(canon) {
+  const enc = new TextEncoder()
+  const buf = await crypto.subtle.digest('SHA-256', enc.encode(canon))
+  const hex = Array.from(new Uint8Array(buf)).map(b => b.toString(16).padStart(2, '0')).join('')
+  return hex.slice(0, 8)
+}
+
+async function onPreview() {
+  busy.value = true
+  try {
+    preview.value = await store.previewMinDim(minW.value, minH.value)
+  } catch (e) {
+    window.__fcToast?.({ text: `Preview failed: ${e.message}`, type: 'error' })
+  } finally {
+    busy.value = false
+  }
+}
+
+async function onDeleteClick() {
+  tokenSha8.value = await sha8(`${minW.value}x${minH.value}`)
+  projectedCounts.value = { 'Images to delete': preview.value.count }
+  showModal.value = true
+}
+
+async function onConfirmedDelete(token) {
+  try {
+    const res = await store.deleteMinDim(minW.value, minH.value, token)
+    window.__fcToast?.({
+      text: `Deleted ${res.deleted} image(s)`, type: 'success',
+    })
+    preview.value = null
+  } catch (e) {
+    window.__fcToast?.({ text: `Delete failed: ${e.message}`, type: 'error' })
+  }
+}
+</script>
+
+<style scoped>
+.fc-clean-card { border-radius: 8px; }
+.fc-muted { color: rgb(var(--v-theme-on-surface-variant)); }
+</style>

frontend/src/components/cleanup/SingleColorAuditCard.vue

@@ -0,0 +1,183 @@
+<template>
+  <v-card class="fc-clean-card">
+    <v-card-title class="d-flex align-center" style="gap: 10px;">
+      <v-icon icon="mdi-palette-swatch" size="small" />
+      <span>Single-color audit</span>
+    </v-card-title>
+    <v-card-text>
+      <p class="fc-muted text-body-2 mb-3">
+        Scan library for images dominated by one color within the
+        tolerance. Catches placeholder / solid-fill / error-page images
+        that slipped through the import filter. Same background-scan
+        cadence as the transparency audit.
+      </p>
+
+      <v-row dense>
+        <v-col cols="6">
+          <v-text-field
+            v-model.number="threshold" label="Threshold (0–1)"
+            type="number" min="0" max="1" step="0.01"
+            density="compact" hide-details
+            :disabled="audit && audit.status === 'running'"
+          />
+        </v-col>
+        <v-col cols="6">
+          <v-text-field
+            v-model.number="tolerance" label="Color tolerance (0–441)"
+            type="number" min="0" max="441"
+            density="compact" hide-details
+            :disabled="audit && audit.status === 'running'"
+          />
+        </v-col>
+      </v-row>
+
+      <v-btn
+        v-if="!audit || audit.status !== 'running'"
+        class="mt-3"
+        color="accent" variant="flat" rounded="pill"
+        prepend-icon="mdi-magnify-scan"
+        :loading="busy"
+        @click="onStart"
+      >Scan library</v-btn>
+
+      <div v-if="audit && audit.status === 'running'" class="mt-3">
+        <v-progress-linear indeterminate color="accent" />
+        <div class="text-body-2 mt-2 d-flex align-center" style="gap: 10px;">
+          <span>
+            Scanning… {{ audit.scanned_count }} checked,
+            {{ audit.matched_count }} matched
+          </span>
+          <v-btn
+            variant="text" size="small" color="warning" rounded="pill"
+            @click="onCancel"
+          >Cancel</v-btn>
+        </div>
+      </div>
+
+      <div v-if="audit && audit.status === 'ready'" class="mt-3">
+        <p class="text-body-2 mb-2">
+          Scan complete. <strong>{{ audit.matched_count }}</strong>
+          image(s) match.
+        </p>
+        <v-btn
+          v-if="audit.matched_count > 0"
+          color="error" variant="flat" rounded="pill"
+          prepend-icon="mdi-delete"
+          @click="onApplyClick"
+        >Delete {{ audit.matched_count }} matching...</v-btn>
+      </div>
+
+      <v-alert
+        v-if="audit && audit.status === 'error'"
+        type="error" variant="tonal" density="compact" class="mt-3"
+      >Scan failed: {{ audit.error }}</v-alert>
+
+      <v-alert
+        v-if="audit && audit.status === 'applied'"
+        type="success" variant="tonal" density="compact" class="mt-3"
+      >Applied — matched images deleted.</v-alert>
+    </v-card-text>
+
+    <DestructiveConfirmModal
+      v-if="audit"
+      v-model="showModal"
+      action="delete"
+      kind="audit"
+      :run-id="audit.id"
+      tier="C"
+      :projected-counts="projectedCounts"
+      description="Permanently deletes images matched by the single-color scan."
+      @confirm="onConfirmedApply"
+    />
+  </v-card>
+</template>
+
+<script setup>
+import { onMounted, onUnmounted, ref } from 'vue'
+
+import DestructiveConfirmModal from '../modal/DestructiveConfirmModal.vue'
+import { useCleanupStore } from '../../stores/cleanup.js'
+
+const store = useCleanupStore()
+const threshold = ref(0.95)
+const tolerance = ref(30)
+const audit = ref(null)
+const busy = ref(false)
+const showModal = ref(false)
+const projectedCounts = ref({})
+let pollTimer = null
+
+onMounted(async () => {
+  await store.loadDefaults()
+  threshold.value = store.defaults.single_color_threshold
+  tolerance.value = store.defaults.single_color_tolerance
+})
+
+onUnmounted(() => stopPoll())
+
+function startPoll(id) {
+  stopPoll()
+  pollTimer = setInterval(async () => {
+    try {
+      const fresh = await store.getAudit(id)
+      audit.value = fresh
+      if (fresh.status !== 'running') stopPoll()
+    } catch (e) {
+      stopPoll()
+      window.__fcToast?.({ text: `Audit poll failed: ${e.message}`, type: 'error' })
+    }
+  }, 5000)
+}
+
+function stopPoll() {
+  if (pollTimer) { clearInterval(pollTimer); pollTimer = null }
+}
+
+async function onStart() {
+  busy.value = true
+  try {
+    const res = await store.startAudit('single_color', {
+      threshold: threshold.value, tolerance: tolerance.value,
+    })
+    audit.value = await store.getAudit(res.audit_id)
+    startPoll(res.audit_id)
+  } catch (e) {
+    window.__fcToast?.({ text: `Scan start failed: ${e.message}`, type: 'error' })
+  } finally {
+    busy.value = false
+  }
+}
+
+async function onCancel() {
+  if (!audit.value) return
+  try {
+    await store.cancelAudit(audit.value.id)
+    audit.value = await store.getAudit(audit.value.id)
+    stopPoll()
+  } catch (e) {
+    window.__fcToast?.({ text: `Cancel failed: ${e.message}`, type: 'error' })
+  }
+}
+
+function onApplyClick() {
+  projectedCounts.value = { 'Images to delete': audit.value.matched_count }
+  showModal.value = true
+}
+
+async function onConfirmedApply(token) {
+  try {
+    const res = await store.applyAudit(audit.value.id, token)
+    window.__fcToast?.({
+      text: `Deleted ${res.deleted} image(s)`, type: 'success',
+    })
+    audit.value = await store.getAudit(audit.value.id)
+  } catch (e) {
+    window.__fcToast?.({ text: `Apply failed: ${e.message}`, type: 'error' })
+  }
+}
+</script>
+
+<style scoped>
+.fc-clean-card { border-radius: 8px; }
+.fc-muted { color: rgb(var(--v-theme-on-surface-variant)); }
+</style>

frontend/src/components/cleanup/TransparencyAuditCard.vue

@@ -0,0 +1,166 @@
+<template>
+  <v-card class="fc-clean-card">
+    <v-card-title class="d-flex align-center" style="gap: 10px;">
+      <v-icon icon="mdi-checkerboard" size="small" />
+      <span>Transparency audit</span>
+    </v-card-title>
+    <v-card-text>
+      <p class="fc-muted text-body-2 mb-3">
+        Scan library for images whose transparent-pixel fraction exceeds
+        the threshold. Animated WebPs / GIFs are skipped (the import-side
+        rule does the same). Runs as a background task — ~50ms per image,
+        so a 57k library takes ~50 minutes.
+      </p>
+
+      <v-text-field
+        v-model.number="threshold" label="Transparency threshold (0–1)"
+        type="number" min="0" max="1" step="0.01" density="compact" hide-details
+        :disabled="audit && audit.status === 'running'"
+        class="mb-3"
+      />
+
+      <v-btn
+        v-if="!audit || audit.status !== 'running'"
+        color="accent" variant="flat" rounded="pill"
+        prepend-icon="mdi-magnify-scan"
+        :loading="busy"
+        @click="onStart"
+      >Scan library</v-btn>
+
+      <div v-if="audit && audit.status === 'running'" class="mt-3">
+        <v-progress-linear indeterminate color="accent" />
+        <div class="text-body-2 mt-2 d-flex align-center" style="gap: 10px;">
+          <span>
+            Scanning… {{ audit.scanned_count }} checked,
+            {{ audit.matched_count }} matched
+          </span>
+          <v-btn
+            variant="text" size="small" color="warning" rounded="pill"
+            @click="onCancel"
+          >Cancel</v-btn>
+        </div>
+      </div>
+
+      <div v-if="audit && audit.status === 'ready'" class="mt-3">
+        <p class="text-body-2 mb-2">
+          Scan complete. <strong>{{ audit.matched_count }}</strong>
+          image(s) match.
+        </p>
+        <v-btn
+          v-if="audit.matched_count > 0"
+          color="error" variant="flat" rounded="pill"
+          prepend-icon="mdi-delete"
+          @click="onApplyClick"
+        >Delete {{ audit.matched_count }} matching...</v-btn>
+      </div>
+
+      <v-alert
+        v-if="audit && audit.status === 'error'"
+        type="error" variant="tonal" density="compact" class="mt-3"
+      >Scan failed: {{ audit.error }}</v-alert>
+
+      <v-alert
+        v-if="audit && audit.status === 'applied'"
+        type="success" variant="tonal" density="compact" class="mt-3"
+      >Applied — matched images deleted.</v-alert>
+    </v-card-text>
+
+    <DestructiveConfirmModal
+      v-if="audit"
+      v-model="showModal"
+      action="delete"
+      kind="audit"
+      :run-id="audit.id"
+      tier="C"
+      :projected-counts="projectedCounts"
+      description="Permanently deletes images matched by the transparency scan."
+      @confirm="onConfirmedApply"
+    />
+  </v-card>
+</template>
+
+<script setup>
+import { onMounted, onUnmounted, ref } from 'vue'
+
+import DestructiveConfirmModal from '../modal/DestructiveConfirmModal.vue'
+import { useCleanupStore } from '../../stores/cleanup.js'
+
+const store = useCleanupStore()
+const threshold = ref(0.9)
+const audit = ref(null)
+const busy = ref(false)
+const showModal = ref(false)
+const projectedCounts = ref({})
+let pollTimer = null
+
+onMounted(async () => {
+  await store.loadDefaults()
+  threshold.value = store.defaults.transparency_threshold
+})
+
+onUnmounted(() => stopPoll())
+
+function startPoll(id) {
+  stopPoll()
+  pollTimer = setInterval(async () => {
+    try {
+      const fresh = await store.getAudit(id)
+      audit.value = fresh
+      if (fresh.status !== 'running') stopPoll()
+    } catch (e) {
+      stopPoll()
+      window.__fcToast?.({ text: `Audit poll failed: ${e.message}`, type: 'error' })
+    }
+  }, 5000)
+}
+
+function stopPoll() {
+  if (pollTimer) { clearInterval(pollTimer); pollTimer = null }
+}
+
+async function onStart() {
+  busy.value = true
+  try {
+    const res = await store.startAudit('transparency', { threshold: threshold.value })
+    audit.value = await store.getAudit(res.audit_id)
+    startPoll(res.audit_id)
+  } catch (e) {
+    window.__fcToast?.({ text: `Scan start failed: ${e.message}`, type: 'error' })
+  } finally {
+    busy.value = false
+  }
+}
+
+async function onCancel() {
+  if (!audit.value) return
+  try {
+    await store.cancelAudit(audit.value.id)
+    audit.value = await store.getAudit(audit.value.id)
+    stopPoll()
+  } catch (e) {
+    window.__fcToast?.({ text: `Cancel failed: ${e.message}`, type: 'error' })
+  }
+}
+
+function onApplyClick() {
+  projectedCounts.value = { 'Images to delete': audit.value.matched_count }
+  showModal.value = true
+}
+
+async function onConfirmedApply(token) {
+  try {
+    const res = await store.applyAudit(audit.value.id, token)
+    window.__fcToast?.({
+      text: `Deleted ${res.deleted} image(s)`, type: 'success',
+    })
+    audit.value = await store.getAudit(audit.value.id)
+  } catch (e) {
+    window.__fcToast?.({ text: `Apply failed: ${e.message}`, type: 'error' })
+  }
+}
+</script>
+
+<style scoped>
+.fc-clean-card { border-radius: 8px; }
+.fc-muted { color: rgb(var(--v-theme-on-surface-variant)); }
+</style>

frontend/src/components/common/ErrorDetailModal.vue

@@ -0,0 +1,87 @@
+<template>
+  <v-dialog :model-value="modelValue" max-width="900"
+            @update:model-value="$emit('update:modelValue', $event)">
+    <v-card>
+      <v-card-title class="d-flex align-center" style="gap: 12px;">
+        <v-icon icon="mdi-alert-circle-outline" color="error" />
+        <span>{{ title }}</span>
+        <v-spacer />
+        <v-btn icon variant="text" size="small" @click="close">
+          <v-icon>mdi-close</v-icon>
+        </v-btn>
+      </v-card-title>
+      <v-card-text>
+        <pre class="fc-err-pre">{{ message || '(no error message)' }}</pre>
+      </v-card-text>
+      <v-card-actions>
+        <v-btn
+          variant="text" rounded="pill" size="small"
+          :prepend-icon="copied ? 'mdi-check' : 'mdi-content-copy'"
+          @click="onCopy"
+        >{{ copied ? 'Copied' : 'Copy' }}</v-btn>
+        <v-spacer />
+        <v-btn variant="text" rounded="pill" @click="close">Close</v-btn>
+      </v-card-actions>
+    </v-card>
+  </v-dialog>
+</template>
+
+<script setup>
+import { ref, watch } from 'vue'
+
+const props = defineProps({
+  modelValue: { type: Boolean, default: false },
+  title: { type: String, default: 'Error details' },
+  message: { type: String, default: '' },
+})
+
+const emit = defineEmits(['update:modelValue'])
+
+const copied = ref(false)
+let copiedTimer = null
+
+watch(() => props.modelValue, (open) => {
+  if (!open) {
+    copied.value = false
+    if (copiedTimer) { clearTimeout(copiedTimer); copiedTimer = null }
+  }
+})
+
+function close() {
+  emit('update:modelValue', false)
+}
+
+async function onCopy() {
+  try {
+    await navigator.clipboard.writeText(props.message || '')
+    copied.value = true
+    if (copiedTimer) clearTimeout(copiedTimer)
+    copiedTimer = setTimeout(() => { copied.value = false }, 1500)
+  } catch (e) {
+    window.__fcToast?.({ text: `Copy failed: ${e.message}`, type: 'error' })
+  }
+}
+</script>
+
+<style scoped>
+/* The full error often contains a SQLAlchemy statement + parameters
+   block + multi-line traceback. Pre-wrap keeps long lines readable;
+   monospace + tabular layout keeps the structure scannable. The
+   max-height + overflow-auto prevents a 50-line traceback from
+   pushing the Close button off-screen. Operator-flagged 2026-05-26:
+   the prior :title="..." tooltip was unusable for content this long. */
+.fc-err-pre {
+  font-family: ui-monospace, SFMono-Regular, Menlo, Consolas, monospace;
+  font-size: 12px;
+  line-height: 1.5;
+  white-space: pre-wrap;
+  word-break: break-word;
+  background: rgb(var(--v-theme-surface-variant, 38 36 41));
+  color: rgb(var(--v-theme-on-surface));
+  padding: 12px 14px;
+  border-radius: 6px;
+  max-height: 60vh;
+  overflow: auto;
+  margin: 0;
+}
+</style>

frontend/src/components/settings/BrowserExtensionCard.vue

@@ -34,11 +34,18 @@
 
       <template v-else-if="manifest?.installed">
         <div class="fc-ext-install mt-3">
+          <!-- Install button: direct :href anchor click (no programmatic
+               window.location.assign). Firefox's XPI-install gesture
+               requires a user-clicked anchor pointing at an
+               application/x-xpinstall response; programmatic navigation
+               sometimes triggered nothing instead of the install dialog
+               (operator-flagged 2026-05-26). No `download` attribute —
+               that would force a save dialog instead of install. -->
           <v-btn
             v-if="isFirefox"
             color="accent" variant="flat" rounded="pill"
             prepend-icon="mdi-firefox"
-            @click="installXpi"
+            :href="manifest.latest_url"
           >Install Firefox extension</v-btn>
 
           <v-btn
@@ -146,11 +153,6 @@ async function loadKey() {
   }
 }
 
-function installXpi() {
-  if (!manifest.value?.latest_url) return
-  window.location.assign(manifest.value.latest_url)
-}
-
 async function rotateKey() {
   rotating.value = true
   try {

frontend/src/components/settings/ImportTaskList.vue

@@ -45,9 +45,11 @@
       <template #item.size_bytes="{ item }">{{ formatBytes(item.size_bytes) }}</template>
       <template #item.created_at="{ item }">{{ formatDate(item.created_at) }}</template>
       <template #item.error="{ item }">
-        <span v-if="item.error" :title="item.error" class="text-caption">
-          {{ shorten(item.error, 60) }}
-        </span>
+        <button
+          v-if="item.error" type="button" class="fc-err-link text-caption"
+          @click="openError(`Task ${item.id} failed`, item.error)"
+          title="Click for full error"
+        >{{ shorten(item.error, 60) }}</button>
       </template>
     </v-data-table-virtual>
     <div v-if="store.hasMore" class="d-flex justify-center py-3">
@@ -100,16 +102,35 @@
         </v-card-actions>
       </v-card>
     </v-dialog>
+
+    <ErrorDetailModal
+      v-model="showErrorModal"
+      :title="errorModalTitle"
+      :message="errorModalMessage"
+    />
   </v-card>
 </template>
 
 <script setup>
 import { computed, ref } from 'vue'
 import { useImportStore } from '../../stores/import.js'
+import ErrorDetailModal from '../common/ErrorDetailModal.vue'
 
 const store = useImportStore()
 const statusFilter = ref(null)
 const clearDialog = ref(false)
+// Click-to-open modal for full error text (operator-flagged 2026-05-26
+// — the prior :title="..." tooltip cramped multi-line SQLAlchemy
+// tracebacks into an unusable popup with no copy-paste affordance).
+const showErrorModal = ref(false)
+const errorModalTitle = ref('')
+const errorModalMessage = ref('')
+
+function openError(title, message) {
+  errorModalTitle.value = title
+  errorModalMessage.value = message || ''
+  showErrorModal.value = true
+}
 const clearAgeDays = ref(7)
 const clearStuckDialog = ref(false)
 
@@ -179,3 +200,21 @@ async function onClearStuckConfirm() {
   clearStuckDialog.value = false
 }
 </script>
+
+<style scoped>
+.fc-err-link {
+  /* Truncated error preview as a clickable button — opens
+     ErrorDetailModal with the full text. Inherits the row's font
+     sizing so it doesn't visually drift from the prior tooltip-bearing
+     span. */
+  color: rgb(var(--v-theme-error, 220 80 80));
+  background: transparent;
+  border: 0;
+  padding: 0;
+  font: inherit;
+  text-align: left;
+  text-decoration: underline dotted;
+  cursor: pointer;
+}
+.fc-err-link:hover { text-decoration: underline; }
+</style>

frontend/src/components/settings/ImportTriggerPanel.vue

@@ -7,8 +7,12 @@
           indeterminate color="accent" size="20"
         />
         <span>
-          Scanning {{ store.activeBatch.source_path }} —
+          {{ store.activeBatch.scan_mode === 'deep' ? 'Deep scanning' : 'Scanning' }}
+          {{ store.activeBatch.source_path || '/import' }} —
           imported {{ store.activeBatch.imported }},
+          <template v-if="store.activeBatch.scan_mode === 'deep'">
+            refreshed {{ store.activeBatch.refreshed || 0 }},
+          </template>
           skipped {{ store.activeBatch.skipped }},
           failed {{ store.activeBatch.failed }} /
           {{ store.activeBatch.total_files }} files
@@ -24,8 +28,13 @@
 
       <p class="text-body-2 mb-3">
         <span v-if="!store.activeBatch">
-          Run a quick scan of the import directory. Deep scan (pHash dedup,
-          archives) lands in FC-2d.
+          <strong>Quick scan</strong> walks <code>/import</code> and enqueues
+          new files only.
+          <strong>Deep scan</strong> additionally re-walks already-imported
+          files so updated sidecar metadata (post title/date/attribution) and
+          previously-NULL phashes / artist links get refreshed. Use after
+          bulk-downloading fresh sidecars for existing content. Both modes
+          route non-media + sidecar pairs through PostAttachment capture.
         </span>
         <span v-else>
           An active batch is in progress. Wait for it to finish, or click
@@ -34,15 +43,26 @@
         </span>
       </p>
 
-      <v-btn
-        color="primary" rounded="pill"
-        :disabled="!!store.activeBatch"
-        :loading="busy"
-        @click="trigger"
-      >
-        <v-icon start>mdi-magnify-scan</v-icon>
-        Quick scan
-      </v-btn>
+      <div class="d-flex flex-wrap" style="gap: 12px;">
+        <v-btn
+          color="primary" rounded="pill"
+          :disabled="!!store.activeBatch"
+          :loading="busy === 'quick'"
+          @click="trigger('quick')"
+        >
+          <v-icon start>mdi-magnify-scan</v-icon>
+          Quick scan
+        </v-btn>
+        <v-btn
+          color="secondary" rounded="pill" variant="tonal"
+          :disabled="!!store.activeBatch"
+          :loading="busy === 'deep'"
+          @click="trigger('deep')"
+        >
+          <v-icon start>mdi-magnify-plus-outline</v-icon>
+          Deep scan
+        </v-btn>
+      </div>
 
       <v-alert v-if="store.triggerError" type="error" variant="tonal" class="mt-3" closable>
         {{ store.triggerError }}
@@ -56,12 +76,12 @@ import { ref } from 'vue'
 import { useImportStore } from '../../stores/import.js'
 
 const store = useImportStore()
-const busy = ref(false)
+const busy = ref(null)
 const clearing = ref(false)
 
-async function trigger() {
-  busy.value = true
-  try { await store.triggerScan() } catch {} finally { busy.value = false }
+async function trigger(mode) {
+  busy.value = mode
+  try { await store.triggerScan(mode) } catch {} finally { busy.value = null }
 }
 
 async function onClearStuck() {

frontend/src/components/settings/MaintenancePanel.vue

@@ -13,8 +13,9 @@
     <AllowlistTable class="mt-4" />
     <AliasTable class="mt-4" />
     <BackupCard class="mt-6" />
-    <TagMaintenanceCard class="mt-6" />
-    <BrowserExtensionCard class="mt-6" />
+    <!-- TagMaintenanceCard moved to Cleanup tab (v26.05.25.7) — it
+         operates on the existing library which fits the Cleanup-tab
+         theme, and clusters with the other audit cards. -->
     <LegacyMigrationCard class="mt-6" />
   </div>
 </template>
@@ -26,8 +27,6 @@ import MLThresholdSliders from './MLThresholdSliders.vue'
 import AllowlistTable from './AllowlistTable.vue'
 import AliasTable from './AliasTable.vue'
 import BackupCard from './BackupCard.vue'
-import TagMaintenanceCard from './TagMaintenanceCard.vue'
-import BrowserExtensionCard from './BrowserExtensionCard.vue'
 import LegacyMigrationCard from './LegacyMigrationCard.vue'
 </script>
 

frontend/src/components/settings/SystemActivityTab.vue

@@ -59,8 +59,12 @@
               <td>{{ r.queue }}</td>
               <td><code>{{ shortTaskName(r.task_name) }}</code></td>
               <td class="fc-tabular">{{ r.target_id ?? '—' }}</td>
-              <td class="fc-err" :title="r.error_message">
-                {{ r.error_type }}
+              <td>
+                <button
+                  type="button" class="fc-err-link"
+                  @click="openError(r.error_type, r.error_message)"
+                  :title="'Click for full error'"
+                >{{ r.error_type }}</button>
               </td>
             </tr>
             <tr v-if="!filteredFailures.length">
@@ -138,6 +142,12 @@
         </div>
       </v-card-text>
     </v-card>
+
+    <ErrorDetailModal
+      v-model="showErrorModal"
+      :title="errorModalTitle"
+      :message="errorModalMessage"
+    />
   </div>
 </template>
 
@@ -145,8 +155,23 @@
 import { computed, onMounted, onUnmounted, ref } from 'vue'
 
 import { useSystemActivityStore } from '../../stores/systemActivity.js'
+import ErrorDetailModal from '../common/ErrorDetailModal.vue'
 import QueuesTable from './QueuesTable.vue'
 
+// Click-to-open modal for full error text. Replaces the unusable
+// :title="..." tooltip (operator-flagged 2026-05-26: SQLAlchemy
+// rollback + traceback content rendered as a cramped browser tooltip
+// you couldn't copy from or scroll within).
+const showErrorModal = ref(false)
+const errorModalTitle = ref('')
+const errorModalMessage = ref('')
+
+function openError(title, message) {
+  errorModalTitle.value = title || 'Error details'
+  errorModalMessage.value = message || ''
+  showErrorModal.value = true
+}
+
 const store = useSystemActivityStore()
 
 const filterQueue = ref(null)
@@ -276,5 +301,18 @@ function formatRelative(iso) {
   font-feature-settings: 'tnum';
 }
 .fc-err { color: rgb(var(--v-theme-error, 220 80 80)); }
+.fc-err-link {
+  /* Styled as a text-only button so the error_type cell stays
+     visually identical to the prior tooltip-bearing row, but is
+     now a real clickable target with hover affordance. */
+  color: rgb(var(--v-theme-error, 220 80 80));
+  background: transparent;
+  border: 0;
+  padding: 0;
+  font: inherit;
+  text-decoration: underline dotted;
+  cursor: pointer;
+}
+.fc-err-link:hover { text-decoration: underline; }
 .fc-muted { color: rgb(var(--v-theme-on-surface-variant)); }
 </style>

frontend/src/stores/cleanup.js

@@ -0,0 +1,72 @@
+import { defineStore } from 'pinia'
+import { ref } from 'vue'
+import { useApi } from '../composables/useApi.js'
+
+export const useCleanupStore = defineStore('cleanup', () => {
+  const api = useApi()
+
+  // Defaults sourced from ImportSettings on mount. Cards pre-fill from
+  // these so the common case ("apply current import filters
+  // retroactively") is one click; operator can override per-audit.
+  const defaults = ref({
+    min_width: 0,
+    min_height: 0,
+    transparency_threshold: 0.9,
+    single_color_threshold: 0.95,
+    single_color_tolerance: 30,
+  })
+
+  const recentRuns = ref([])
+
+  async function loadDefaults() {
+    const s = await api.get('/api/settings/import')
+    defaults.value = {
+      min_width: s.min_width ?? 0,
+      min_height: s.min_height ?? 0,
+      transparency_threshold: s.transparency_threshold ?? 0.9,
+      single_color_threshold: s.single_color_threshold ?? 0.95,
+      single_color_tolerance: s.single_color_tolerance ?? 30,
+    }
+  }
+
+  async function previewMinDim(min_width, min_height) {
+    return await api.post('/api/cleanup/min-dimension/preview', {
+      body: { min_width, min_height },
+    })
+  }
+
+  async function deleteMinDim(min_width, min_height, confirm) {
+    return await api.post('/api/cleanup/min-dimension/delete', {
+      body: { min_width, min_height, confirm },
+    })
+  }
+
+  async function startAudit(rule, params) {
+    return await api.post('/api/cleanup/audit', { body: { rule, params } })
+  }
+
+  async function getAudit(id) {
+    return await api.get(`/api/cleanup/audit/${id}`)
+  }
+
+  async function loadHistory(limit = 20) {
+    const body = await api.get(`/api/cleanup/audit?limit=${limit}`)
+    recentRuns.value = body.runs
+    return body.runs
+  }
+
+  async function applyAudit(id, confirm) {
+    return await api.post(`/api/cleanup/audit/${id}/apply`, { body: { confirm } })
+  }
+
+  async function cancelAudit(id) {
+    return await api.post(`/api/cleanup/audit/${id}/cancel`)
+  }
+
+  return {
+    defaults, recentRuns,
+    loadDefaults,
+    previewMinDim, deleteMinDim,
+    startAudit, getAudit, loadHistory, applyAudit, cancelAudit,
+  }
+})

frontend/src/stores/import.js

@@ -52,11 +52,58 @@ export const useImportStore = defineStore('import', () => {
     }
   }
 
-  async function triggerScan() {
+  async function triggerScan(mode = 'quick') {
+    if (!['quick', 'deep', 'verify'].includes(mode)) {
+      throw new Error(`unsupported scan mode: ${mode}`)
+    }
     triggerError.value = null
     try {
-      await api.post('/api/import/trigger', { body: { mode: 'quick' } })
+      await api.post('/api/import/trigger', { body: { mode } })
+      // Acknowledge immediately so the click isn't invisible. scan_directory
+      // can finalize the batch synchronously when every file in /import is
+      // already on a non-failed ImportTask (operator-flagged 2026-05-25:
+      // 233k existing tasks → all paths in skip-set → files_seen=0 →
+      // batch flashes 'running' for <100ms then 'complete' before the
+      // first refreshStatus() lands; UI never sees the active state).
+      const label = mode === 'deep'
+        ? 'Deep scan triggered (re-applying sidecar metadata + filling NULL phash/artist on existing rows)'
+        : mode === 'verify' ? 'Library verify triggered' : 'Quick scan triggered'
+      window.__fcToast?.({ text: label, type: 'success' })
       await refreshStatus()
+      // Re-poll twice over ~5s and produce an HONEST follow-up toast.
+      // Operator-flagged 2026-05-25: the prior "no new files" message was
+      // misleading because deep scan IS doing work (refresh) even when
+      // there are no new files to import. Surface the real workload count
+      // (imported + refreshed + queued) instead. For quick scan + zero
+      // queued work, fall back to "up to date" instead of the old
+      // implementation-detail-leaking message.
+      setTimeout(async () => {
+        await refreshStatus()
+        await loadTasks(true)
+        if (activeBatch.value || mode === 'verify') return
+        // Batch finalized quickly; figure out what actually happened.
+        // The task list was just refreshed; the freshest row(s) carry
+        // the batch outcome.
+        const batchId = tasks.value[0]?.batch_id
+        const sameBatch = batchId
+          ? tasks.value.filter(t => t.batch_id === batchId)
+          : []
+        const refreshedCount = sameBatch.filter(t => t.status === 'complete' && t.result_image_id).length
+        const newImported = sameBatch.filter(t => t.status === 'complete' && t.result_image_id && !t.error).length
+        if (mode === 'deep' && sameBatch.length > 0) {
+          window.__fcToast?.({
+            text: `Deep scan finished — ${sameBatch.length} file(s) processed`,
+            type: 'info',
+          })
+        } else if (mode === 'quick' && sameBatch.length > 0) {
+          window.__fcToast?.({
+            text: `Quick scan finished — ${newImported} new file(s) queued`,
+            type: 'info',
+          })
+        } else {
+          window.__fcToast?.({ text: 'Library is up to date', type: 'info' })
+        }
+      }, 2000)
     } catch (e) {
       triggerError.value = e.message
       window.__fcToast?.({ text: `Scan failed: ${e.message}`, type: 'error' })

frontend/src/views/ArtistView.vue

@@ -45,69 +45,92 @@
         >Credential health · FC-3b</v-chip>
       </div>
 
-      <section v-if="store.overview.cooccurring_tags.length" class="fc-artist__sec">
-        <h2 class="fc-h2">Frequent tags</h2>
-        <div class="fc-artist__tags">
-          <v-chip
-            v-for="t in store.overview.cooccurring_tags" :key="t.id"
-            size="small" @click="openTag(t.id)"
-          >{{ t.name }} <span class="fc-artist__tagc">{{ t.count }}</span></v-chip>
-        </div>
-      </section>
+      <!-- Tabs split (2026-05-25): Settings was previously slotted at the
+           bottom of the page after the infinite-scroll image grid, which
+           made it effectively unreachable for any artist with more than
+           a couple of pages of content. The Settings tab now hosts
+           destructive admin actions (artist+content cascade-delete) and
+           any future per-artist management UI. v-tabs is `position:
+           sticky; top: 64px` (under the 64px AppShell TopNav) so it
+           stays parked while the gallery scrolls. -->
+      <v-tabs
+        v-model="tab" color="accent" class="mb-4"
+        style="position: sticky; top: 64px; z-index: 4;
+               background: rgb(var(--v-theme-surface));"
+      >
+        <v-tab value="overview">Overview</v-tab>
+        <v-tab value="settings">Settings</v-tab>
+      </v-tabs>
 
-      <section v-if="store.overview.activity.length" class="fc-artist__sec">
-        <h2 class="fc-h2">Activity</h2>
-        <svg class="fc-artist__spark" :viewBox="`0 0 ${sparkW} ${sparkH}`"
-             preserveAspectRatio="none" role="img" aria-label="posts over time">
-          <polyline :points="sparkPoints" fill="none"
-                    stroke="rgb(var(--v-theme-accent))" stroke-width="2" />
-        </svg>
-      </section>
+      <v-window v-model="tab">
+        <v-window-item value="overview">
+          <section v-if="store.overview.cooccurring_tags.length" class="fc-artist__sec">
+            <h2 class="fc-h2">Frequent tags</h2>
+            <div class="fc-artist__tags">
+              <v-chip
+                v-for="t in store.overview.cooccurring_tags" :key="t.id"
+                size="small" @click="openTag(t.id)"
+              >{{ t.name }} <span class="fc-artist__tagc">{{ t.count }}</span></v-chip>
+            </div>
+          </section>
 
-      <section v-if="store.overview.sources.length" class="fc-artist__sec">
-        <div class="fc-artist__sec-head">
-          <h2 class="fc-h2">Sources</h2>
-          <RouterLink
-            :to="`/subscriptions?artist_id=${store.overview.id}`"
-            class="fc-artist__manage"
-          >Manage subscriptions →</RouterLink>
-        </div>
-        <v-table density="compact">
-          <thead>
-            <tr><th>Platform</th><th>URL</th><th class="text-right">Images</th></tr>
-          </thead>
-          <tbody>
-            <tr v-for="s in store.overview.sources" :key="s.id">
-              <td>{{ s.platform }}</td>
-              <td class="fc-artist__url">{{ s.url }}</td>
-              <td class="text-right">{{ s.image_count }}</td>
-            </tr>
-          </tbody>
-        </v-table>
-      </section>
+          <section v-if="store.overview.activity.length" class="fc-artist__sec">
+            <h2 class="fc-h2">Activity</h2>
+            <svg class="fc-artist__spark" :viewBox="`0 0 ${sparkW} ${sparkH}`"
+                 preserveAspectRatio="none" role="img" aria-label="posts over time">
+              <polyline :points="sparkPoints" fill="none"
+                        stroke="rgb(var(--v-theme-accent))" stroke-width="2" />
+            </svg>
+          </section>
 
-      <section class="fc-artist__sec">
-        <h2 class="fc-h2">Images</h2>
-        <MasonryGrid
-          :items="store.images"
-          :loading="store.imagesLoading"
-          :has-more="store.hasMoreImages"
-          @load-more="store.loadMoreImages(slug)"
-          @open="openImage"
-        />
-      </section>
+          <section v-if="store.overview.sources.length" class="fc-artist__sec">
+            <div class="fc-artist__sec-head">
+              <h2 class="fc-h2">Sources</h2>
+              <RouterLink
+                :to="`/subscriptions?artist_id=${store.overview.id}`"
+                class="fc-artist__manage"
+              >Manage subscriptions →</RouterLink>
+            </div>
+            <v-table density="compact">
+              <thead>
+                <tr><th>Platform</th><th>URL</th><th class="text-right">Images</th></tr>
+              </thead>
+              <tbody>
+                <tr v-for="s in store.overview.sources" :key="s.id">
+                  <td>{{ s.platform }}</td>
+                  <td class="fc-artist__url">{{ s.url }}</td>
+                  <td class="text-right">{{ s.image_count }}</td>
+                </tr>
+              </tbody>
+            </v-table>
+          </section>
 
-      <ArtistDangerZone
-        :slug="slug"
-        :artist-id="store.overview.id"
-        :artist-name="store.overview.name"
-      />
+          <section class="fc-artist__sec">
+            <h2 class="fc-h2">Images</h2>
+            <MasonryGrid
+              :items="store.images"
+              :loading="store.imagesLoading"
+              :has-more="store.hasMoreImages"
+              @load-more="store.loadMoreImages(slug)"
+              @open="openImage"
+            />
+          </section>
+        </v-window-item>
+
+        <v-window-item value="settings">
+          <ArtistDangerZone
+            :slug="slug"
+            :artist-id="store.overview.id"
+            :artist-name="store.overview.name"
+          />
+        </v-window-item>
+      </v-window>
     </template>
   </v-container>
 </template>
 
 <script setup>
-import { computed, watch } from 'vue'
+import { computed, ref, watch } from 'vue'
 import { useRoute, useRouter, RouterLink } from 'vue-router'
 import { useArtistStore } from '../stores/artist.js'
 import { useModalStore } from '../stores/modal.js'
@@ -120,8 +143,18 @@ const store = useArtistStore()
 const modal = useModalStore()
 
 const slug = computed(() => route.params.slug)
+// Per-artist tab — defaults to Overview. Settings tab hosts destructive
+// admin actions (DangerZone). Switching artists resets to Overview so the
+// destructive surface isn't re-shown by accident when navigating between
+// artists.
+const tab = ref('overview')
 
-watch(slug, (s) => { if (s) store.load(s) }, { immediate: true })
+watch(slug, (s) => {
+  if (s) {
+    store.load(s)
+    tab.value = 'overview'
+  }
+}, { immediate: true })
 
 const dateRange = computed(() => {
   const r = store.overview?.date_range

frontend/src/views/CleanupView.vue

@@ -0,0 +1,31 @@
+<template>
+  <div class="fc-cleanup">
+    <p class="fc-muted text-body-2 mb-4">
+      Retroactive enforcement of import-filter rules. Each card scans the
+      existing library for content that the current import filters would
+      now exclude. Destructive — typed-token confirmation required.
+    </p>
+
+    <MinDimensionCard class="mb-4" />
+    <TransparencyAuditCard class="mb-4" />
+    <SingleColorAuditCard class="mb-4" />
+    <TagMaintenanceCard />
+  </div>
+</template>
+
+<script setup>
+import MinDimensionCard from '../components/cleanup/MinDimensionCard.vue'
+import TransparencyAuditCard from '../components/cleanup/TransparencyAuditCard.vue'
+import SingleColorAuditCard from '../components/cleanup/SingleColorAuditCard.vue'
+// Reuse existing TagMaintenanceCard (FC-3k) as-is — it already handles
+// preview + commit of prune-unused-tags via the admin store. Operator
+// confirmed 2026-05-26: don't duplicate into a new UnusedTagsCard.
+// MaintenancePanel drops its TagMaintenanceCard reference in the
+// SettingsView edit (Task 16) so this is now the sole rendering site.
+import TagMaintenanceCard from '../components/settings/TagMaintenanceCard.vue'
+</script>
+
+<style scoped>
+.fc-cleanup { max-width: 900px; }
+.fc-muted { color: rgb(var(--v-theme-on-surface-variant)); }
+</style>

frontend/src/views/SettingsView.vue

@@ -1,9 +1,20 @@
 <template>
   <v-container fluid class="py-6">
-    <v-tabs v-model="tab" color="accent" class="mb-4">
+    <!-- Sticky tabs: operator-flagged 2026-05-25 — long Import / Maintenance
+         panels pushed the tab strip out of the viewport, forcing a scroll-
+         to-top just to change tab. AppShell's TopNav is 64px sticky, so the
+         tab strip lives directly under it. Background uses the theme surface
+         token so it visually merges with the page rather than the
+         translucent v-tabs default. -->
+    <v-tabs
+      v-model="tab" color="accent" class="mb-4"
+      style="position: sticky; top: 64px; z-index: 4;
+             background: rgb(var(--v-theme-surface));"
+    >
       <v-tab value="overview">Overview</v-tab>
       <v-tab value="activity">Activity</v-tab>
       <v-tab value="import">Import</v-tab>
+      <v-tab value="cleanup">Cleanup</v-tab>
       <v-tab value="maintenance">Maintenance</v-tab>
     </v-tabs>
 
@@ -21,6 +32,11 @@
           {{ system.stats.tasks.pending + system.stats.tasks.queued }} import task(s) pending.
           <v-btn variant="text" size="small" @click="tab = 'import'">Go to Import tab</v-btn>
         </v-alert>
+        <!-- Browser-extension install/download lives on Overview (moved
+             from Maintenance 2026-05-25). Overview is the discovery
+             surface for "things to set up"; Maintenance is for
+             housekeeping of already-set-up systems. -->
+        <BrowserExtensionCard class="mt-6" />
       </v-window-item>
 
       <v-window-item value="activity">
@@ -28,11 +44,18 @@
       </v-window-item>
 
       <v-window-item value="import">
+        <!-- Order: trigger → recent tasks → filters. Tasks sit directly
+             below the trigger so operator sees hit/miss feedback without
+             scrolling past the filter card (operator-flagged 2026-05-25). -->
         <ImportTriggerPanel />
         <v-divider class="my-6" />
-        <ImportFiltersForm />
-        <v-divider class="my-6" />
         <ImportTaskList />
+        <v-divider class="my-6" />
+        <ImportFiltersForm />
+      </v-window-item>
+
+      <v-window-item value="cleanup">
+        <CleanupView />
       </v-window-item>
 
       <v-window-item value="maintenance">
@@ -49,10 +72,12 @@ import { useImportStore } from '../stores/import.js'
 import SystemStatsCards from '../components/settings/SystemStatsCards.vue'
 import SystemActivitySummary from '../components/settings/SystemActivitySummary.vue'
 import SystemActivityTab from '../components/settings/SystemActivityTab.vue'
+import BrowserExtensionCard from '../components/settings/BrowserExtensionCard.vue'
 import ImportTriggerPanel from '../components/settings/ImportTriggerPanel.vue'
 import ImportFiltersForm from '../components/settings/ImportFiltersForm.vue'
 import ImportTaskList from '../components/settings/ImportTaskList.vue'
 import MaintenancePanel from '../components/settings/MaintenancePanel.vue'
+import CleanupView from './CleanupView.vue'
 import { useMLStore } from '../stores/ml.js'
 
 const tab = ref('overview')

tests/test_api_cleanup.py

@@ -0,0 +1,215 @@
+"""API tests for the /api/cleanup/* blueprint.
+
+Per reference-async-coredml-test-assertions, post-DML state checks go
+via column selects, not ORM entity access.
+"""
+
+import hashlib
+from datetime import UTC, datetime
+
+import pytest
+from PIL import Image
+from sqlalchemy import func, select
+
+from backend.app import create_app
+from backend.app.celery_app import celery
+from backend.app.models import ImageRecord, LibraryAuditRun
+
+pytestmark = pytest.mark.integration
+
+
+@pytest.fixture(autouse=True)
+def disable_celery_eager(monkeypatch):
+    monkeypatch.setattr(celery.conf, "task_always_eager", False)
+
+
+@pytest.fixture
+async def app():
+    return create_app()
+
+
+@pytest.fixture
+async def client(app):
+    async with app.test_client() as c:
+        yield c
+
+
+def _sha256_min_dim_token(min_w: int, min_h: int) -> str:
+    canon = f"{min_w}x{min_h}"
+    return f"delete-min-dim-{hashlib.sha256(canon.encode()).hexdigest()[:8]}"
+
+
+async def _seed_image(db, tmp_path, *, w, h, name):
+    path = tmp_path / name
+    Image.new("RGB", (w, h), (w % 256, h % 256, 0)).save(path)
+    sha = f"api-cleanup-{name}".ljust(64, "x")[:64]
+    rec = ImageRecord(
+        path=str(path), sha256=sha,
+        size_bytes=path.stat().st_size, mime="image/png",
+        width=w, height=h, origin="imported_filesystem",
+        integrity_status="ok",
+    )
+    db.add(rec)
+    await db.flush()
+    return rec
+
+
+@pytest.mark.asyncio
+async def test_min_dimension_preview_returns_count(client, db, tmp_path):
+    await _seed_image(db, tmp_path, w=50, h=50, name="small.png")
+    await _seed_image(db, tmp_path, w=500, h=500, name="big.png")
+    await db.commit()
+    resp = await client.post(
+        "/api/cleanup/min-dimension/preview",
+        json={"min_width": 200, "min_height": 200},
+    )
+    assert resp.status_code == 200
+    body = await resp.get_json()
+    assert body["count"] == 1
+
+
+@pytest.mark.asyncio
+async def test_min_dimension_delete_with_token_removes_rows(client, db, tmp_path):
+    await _seed_image(db, tmp_path, w=50, h=50, name="s2.png")
+    await _seed_image(db, tmp_path, w=500, h=500, name="b2.png")
+    await db.commit()
+    token = _sha256_min_dim_token(200, 200)
+    resp = await client.post(
+        "/api/cleanup/min-dimension/delete",
+        json={"min_width": 200, "min_height": 200, "confirm": token},
+    )
+    assert resp.status_code == 200
+    body = await resp.get_json()
+    assert body["deleted"] == 1
+    remaining = await db.execute(select(func.count()).select_from(ImageRecord))
+    assert remaining.scalar_one() == 1
+
+
+@pytest.mark.asyncio
+async def test_min_dimension_delete_with_bad_token_returns_400(client, db, tmp_path):
+    await _seed_image(db, tmp_path, w=50, h=50, name="s3.png")
+    await db.commit()
+    resp = await client.post(
+        "/api/cleanup/min-dimension/delete",
+        json={"min_width": 200, "min_height": 200, "confirm": "nope"},
+    )
+    assert resp.status_code == 400
+    body = await resp.get_json()
+    assert body["error"] == "confirm_mismatch"
+
+
+@pytest.mark.asyncio
+async def test_audit_create_returns_id_and_running_status(
+    client, db, monkeypatch,
+):
+    from backend.app.tasks import library_audit
+    monkeypatch.setattr(
+        library_audit.scan_library_for_rule, "delay", lambda audit_id: None,
+    )
+    resp = await client.post(
+        "/api/cleanup/audit",
+        json={"rule": "transparency", "params": {"threshold": 0.9}},
+    )
+    assert resp.status_code == 202
+    body = await resp.get_json()
+    assert body["status"] == "running"
+    assert isinstance(body["audit_id"], int)
+
+
+@pytest.mark.asyncio
+async def test_audit_create_returns_409_when_another_is_running(client, db):
+    audit = LibraryAuditRun(
+        rule="transparency", params={"threshold": 0.9},
+        status="running", matched_ids=[],
+    )
+    db.add(audit)
+    await db.commit()
+    resp = await client.post(
+        "/api/cleanup/audit",
+        json={"rule": "single_color", "params": {"threshold": 0.95, "tolerance": 30}},
+    )
+    assert resp.status_code == 409
+
+
+@pytest.mark.asyncio
+async def test_audit_get_by_id_returns_full_row(client, db):
+    audit = LibraryAuditRun(
+        rule="transparency", params={"threshold": 0.85},
+        status="ready", scanned_count=100, matched_count=3,
+        matched_ids=[1, 2, 3],
+    )
+    db.add(audit)
+    await db.commit()
+    resp = await client.get(f"/api/cleanup/audit/{audit.id}")
+    assert resp.status_code == 200
+    body = await resp.get_json()
+    assert body["rule"] == "transparency"
+    assert body["matched_count"] == 3
+    assert body["matched_ids"] == [1, 2, 3]
+
+
+@pytest.mark.asyncio
+async def test_audit_history_returns_recent_runs(client, db):
+    for _ in range(3):
+        db.add(LibraryAuditRun(
+            rule="transparency", params={"threshold": 0.9},
+            status="applied", matched_ids=[],
+            finished_at=datetime.now(UTC),
+        ))
+    await db.commit()
+    resp = await client.get("/api/cleanup/audit?limit=5")
+    assert resp.status_code == 200
+    body = await resp.get_json()
+    assert len(body["runs"]) >= 3
+
+
+@pytest.mark.asyncio
+async def test_audit_apply_with_token_deletes(client, db, tmp_path):
+    rec = await _seed_image(db, tmp_path, w=100, h=100, name="apply.png")
+    audit = LibraryAuditRun(
+        rule="transparency", params={"threshold": 0.9},
+        status="ready", scanned_count=1, matched_count=1,
+        matched_ids=[rec.id],
+    )
+    db.add(audit)
+    await db.commit()
+    resp = await client.post(
+        f"/api/cleanup/audit/{audit.id}/apply",
+        json={"confirm": f"delete-audit-{audit.id}"},
+    )
+    assert resp.status_code == 200
+    body = await resp.get_json()
+    assert body["deleted"] == 1
+
+
+@pytest.mark.asyncio
+async def test_audit_apply_with_bad_token_returns_400(client, db):
+    audit = LibraryAuditRun(
+        rule="transparency", params={"threshold": 0.9},
+        status="ready", matched_ids=[],
+    )
+    db.add(audit)
+    await db.commit()
+    resp = await client.post(
+        f"/api/cleanup/audit/{audit.id}/apply",
+        json={"confirm": "wrong-token"},
+    )
+    assert resp.status_code == 400
+    body = await resp.get_json()
+    assert body["error"] == "confirm_mismatch"
+
+
+@pytest.mark.asyncio
+async def test_audit_cancel_flips_status(client, db):
+    audit = LibraryAuditRun(
+        rule="transparency", params={"threshold": 0.9},
+        status="running", matched_ids=[],
+    )
+    db.add(audit)
+    await db.commit()
+    resp = await client.post(f"/api/cleanup/audit/{audit.id}/cancel")
+    assert resp.status_code == 200
+    new_status = await db.execute(
+        select(LibraryAuditRun.status).where(LibraryAuditRun.id == audit.id)
+    )
+    assert new_status.scalar_one() == "cancelled"

tests/test_audits_single_color.py

@@ -0,0 +1,42 @@
+"""Tests for the single-color audit rule.
+
+The rule downsamples + measures the fraction of pixels within `tolerance`
+(Euclidean RGB distance) of the dominant color. Matches if that fraction
+exceeds `threshold`. Single-color content is typically uploaded by
+mistake (placeholder/error/preview images) and should be flagged.
+"""
+
+from PIL import Image
+
+from backend.app.services.audits import single_color
+
+
+def test_single_color_evaluate_true_for_uniform_image():
+    im = Image.new("RGB", (50, 50), (128, 64, 200))
+    assert single_color.evaluate(im, threshold=0.9, tolerance=10) is True
+
+
+def test_single_color_evaluate_false_for_diverse_image():
+    # Half black, half white — no single color dominates.
+    im = Image.new("RGB", (50, 50), (0, 0, 0))
+    for x in range(25):
+        for y in range(50):
+            im.putpixel((x, y), (255, 255, 255))
+    assert single_color.evaluate(im, threshold=0.9, tolerance=10) is False
+
+
+def test_single_color_evaluate_respects_tolerance_widening():
+    # Gradient image: pixels span 0..50 in R channel. Tight tolerance
+    # rejects (no concentration), wide tolerance accepts (all near 25).
+    im = Image.new("RGB", (50, 50), (0, 0, 0))
+    for x in range(50):
+        for y in range(50):
+            im.putpixel((x, y), (x, 0, 0))
+    assert single_color.evaluate(im, threshold=0.9, tolerance=5) is False
+    assert single_color.evaluate(im, threshold=0.9, tolerance=50) is True
+
+
+def test_single_color_evaluate_handles_rgba_input():
+    # Alpha channel should be ignored — only RGB matters for the rule.
+    im = Image.new("RGBA", (50, 50), (100, 100, 100, 128))
+    assert single_color.evaluate(im, threshold=0.9, tolerance=10) is True

tests/test_audits_transparency.py

@@ -0,0 +1,46 @@
+"""Tests for the transparency audit rule.
+
+The rule mirrors `Importer._transparency_pct` semantics for retroactive
+enforcement: returns True iff the fraction of fully-transparent pixels
+exceeds the threshold. Animated images short-circuit to False to avoid
+the multi-frame PIL decode that triggered SoftTimeLimitExceeded
+2026-05-26 against animated WebPs.
+"""
+
+from PIL import Image
+
+from backend.app.services.audits import transparency
+
+
+def test_transparency_evaluate_true_when_fully_transparent():
+    im = Image.new("RGBA", (10, 10), (0, 0, 0, 0))
+    assert transparency.evaluate(im, threshold=0.5) is True
+
+
+def test_transparency_evaluate_false_when_fully_opaque():
+    im = Image.new("RGBA", (10, 10), (200, 100, 50, 255))
+    assert transparency.evaluate(im, threshold=0.5) is False
+
+
+def test_transparency_evaluate_respects_threshold_boundary():
+    # Half-transparent image: 50% alpha=0 pixels, 50% alpha=255.
+    im = Image.new("RGBA", (10, 10), (0, 0, 0, 0))
+    for x in range(5):
+        for y in range(10):
+            im.putpixel((x, y), (0, 0, 0, 255))
+    # 50% transparent. threshold=0.4 → True; threshold=0.6 → False.
+    assert transparency.evaluate(im, threshold=0.4) is True
+    assert transparency.evaluate(im, threshold=0.6) is False
+
+
+def test_transparency_evaluate_false_for_rgb_image_without_alpha():
+    im = Image.new("RGB", (10, 10), (128, 128, 128))
+    assert transparency.evaluate(im, threshold=0.5) is False
+
+
+def test_transparency_evaluate_false_for_animated_image():
+    im = Image.new("RGBA", (10, 10), (0, 0, 0, 0))
+    # Mark as animated (mimics PIL's WebP/GIF multi-frame attribute).
+    im.is_animated = True  # type: ignore[attr-defined]
+    im.n_frames = 5  # type: ignore[attr-defined]
+    assert transparency.evaluate(im, threshold=0.5) is False

tests/test_cleanup_service_audit.py

@@ -0,0 +1,175 @@
+"""Tests for cleanup_service's library-audit additions.
+
+Covers:
+  - project_min_dimension_violations: SQL-only query against width/height
+  - delete_min_dimension_violations: routes through existing delete_images
+  - audit lifecycle (start_audit_run / apply_audit_run / cancel_audit_run)
+
+Tests assert via column selects per reference-async-coredml-test-assertions
+(post-DML ORM entity access via session.get() raises MissingGreenlet on
+async sessions; we use db_sync here but the convention is preserved for
+consistency).
+"""
+
+from pathlib import Path
+
+import pytest
+from PIL import Image
+from sqlalchemy import func, select
+
+from backend.app.models import ImageRecord, LibraryAuditRun
+from backend.app.services import cleanup_service
+
+pytestmark = pytest.mark.integration
+
+
+def _make_image_record(db_sync, tmp_path, *, width, height, color):
+    """Helper: write a real PIL file to a UNIQUE path (reference-image-record-path-unique)
+    and insert an ImageRecord row with all required NOT-NULL columns set."""
+    path = tmp_path / f"img_{width}x{height}_{color[0]}.png"
+    Image.new("RGB", (width, height), color).save(path)
+    # sha256 column is varchar(64) — use a deterministic 64-char pseudo
+    # hash built from the unique inputs.
+    sha = f"{width:04d}{height:04d}{color[0]:03d}".ljust(64, "0")[:64]
+    rec = ImageRecord(
+        path=str(path),
+        sha256=sha,
+        size_bytes=path.stat().st_size,
+        mime="image/png",  # reference-image-record-required-columns
+        width=width,
+        height=height,
+        origin="imported_filesystem",  # feedback-check-existing-enums
+        integrity_status="ok",
+    )
+    db_sync.add(rec)
+    db_sync.flush()
+    return rec
+
+
+def test_project_min_dimension_violations_returns_count_and_samples(db_sync, tmp_path):
+    _make_image_record(db_sync, tmp_path, width=100, height=100, color=(10, 0, 0))
+    _make_image_record(db_sync, tmp_path, width=50, height=50, color=(20, 0, 0))
+    _make_image_record(db_sync, tmp_path, width=400, height=400, color=(30, 0, 0))
+    db_sync.commit()
+
+    result = cleanup_service.project_min_dimension_violations(
+        db_sync, min_width=200, min_height=200,
+    )
+    assert result["count"] == 2  # 100x100 and 50x50 violate
+    assert len(result["sample_ids"]) == 2
+
+
+def test_delete_min_dimension_violations_unlinks_and_cascades(db_sync, tmp_path):
+    rec_small = _make_image_record(db_sync, tmp_path, width=50, height=50, color=(40, 0, 0))
+    rec_big = _make_image_record(db_sync, tmp_path, width=500, height=500, color=(50, 0, 0))
+    small_path = Path(rec_small.path)
+    db_sync.commit()
+
+    # images_root is tmp_path here because the test fixtures stored files there;
+    # delete_images() uses it to unlink originals + thumbs from the right tree.
+    deleted = cleanup_service.delete_min_dimension_violations(
+        db_sync, min_width=200, min_height=200, images_root=tmp_path,
+    )
+    assert deleted == 1
+    # Verify via column selects per banked rule.
+    remaining_ids = db_sync.execute(
+        select(ImageRecord.id).order_by(ImageRecord.id)
+    ).scalars().all()
+    assert remaining_ids == [rec_big.id]
+    assert not small_path.exists()  # file unlinked
+
+
+# --- Audit lifecycle tests (Task 5) ---
+
+import backend.app.tasks.library_audit  # noqa: F401, E402 — celery registration
+
+
+def test_start_audit_run_creates_row_and_dispatches(db_sync, monkeypatch):
+    dispatched = []
+    from backend.app.tasks import library_audit as la_mod
+    monkeypatch.setattr(
+        la_mod.scan_library_for_rule, "delay",
+        lambda audit_id: dispatched.append(audit_id),
+    )
+    audit_id = cleanup_service.start_audit_run(
+        db_sync, rule="transparency", params={"threshold": 0.9},
+    )
+    db_sync.commit()
+    row_status = db_sync.execute(
+        select(LibraryAuditRun.status).where(LibraryAuditRun.id == audit_id)
+    ).scalar_one()
+    assert row_status == "running"
+    assert dispatched == [audit_id]
+
+
+def test_start_audit_run_rejects_when_another_is_running(db_sync, monkeypatch):
+    from backend.app.tasks import library_audit as la_mod
+    monkeypatch.setattr(
+        la_mod.scan_library_for_rule, "delay", lambda audit_id: None,
+    )
+    cleanup_service.start_audit_run(
+        db_sync, rule="transparency", params={"threshold": 0.9},
+    )
+    db_sync.commit()
+    with pytest.raises(cleanup_service.AuditAlreadyRunning):
+        cleanup_service.start_audit_run(
+            db_sync, rule="single_color",
+            params={"threshold": 0.95, "tolerance": 30},
+        )
+
+
+def test_apply_audit_run_with_correct_token_deletes_matched(db_sync, tmp_path):
+    rec = _make_image_record(
+        db_sync, tmp_path, width=100, height=100, color=(60, 0, 0),
+    )
+    db_sync.flush()
+    audit = LibraryAuditRun(
+        rule="transparency", params={"threshold": 0.9},
+        status="ready", scanned_count=1, matched_count=1,
+        matched_ids=[rec.id],
+    )
+    db_sync.add(audit)
+    db_sync.commit()
+
+    deleted = cleanup_service.apply_audit_run(
+        db_sync, audit_id=audit.id,
+        confirm_token=f"delete-audit-{audit.id}",
+        images_root=tmp_path,
+    )
+    assert deleted == 1
+    remaining_count = db_sync.execute(
+        select(func.count()).select_from(ImageRecord)
+    ).scalar_one()
+    assert remaining_count == 0
+    new_status = db_sync.execute(
+        select(LibraryAuditRun.status).where(LibraryAuditRun.id == audit.id)
+    ).scalar_one()
+    assert new_status == "applied"
+
+
+def test_apply_audit_run_with_wrong_token_raises(db_sync, tmp_path):
+    audit = LibraryAuditRun(
+        rule="transparency", params={"threshold": 0.9},
+        status="ready", matched_ids=[],
+    )
+    db_sync.add(audit)
+    db_sync.commit()
+    with pytest.raises(cleanup_service.ConfirmTokenMismatch):
+        cleanup_service.apply_audit_run(
+            db_sync, audit_id=audit.id,
+            confirm_token="wrong-token", images_root=tmp_path,
+        )
+
+
+def test_cancel_audit_run_flips_status(db_sync):
+    audit = LibraryAuditRun(
+        rule="transparency", params={"threshold": 0.9},
+        status="running", matched_ids=[],
+    )
+    db_sync.add(audit)
+    db_sync.commit()
+    cleanup_service.cancel_audit_run(db_sync, audit_id=audit.id)
+    new_status = db_sync.execute(
+        select(LibraryAuditRun.status).where(LibraryAuditRun.id == audit.id)
+    ).scalar_one()
+    assert new_status == "cancelled"

tests/test_importer_attachments.py

@@ -59,3 +59,25 @@ def test_json_sidecar_is_not_attached(importer, import_layout):
         select(func.count()).select_from(PostAttachment)
     ).scalar_one()
     assert n == 0
+
+
+def test_mangled_filename_extension_is_sanitized(importer, import_layout):
+    """gallery-dl sometimes URL-encodes a query string into the basename
+    (`...https___www.patreon.com_media-u_Z0F...`). Python's Path.suffix
+    returns 50+ chars of base64-ish junk for those, which blows the
+    PostAttachment.ext varchar(32) column. Operator-flagged 2026-05-25.
+    The importer should record an empty ext rather than crash."""
+    import_root, _ = import_layout
+    f = (
+        import_root / "Alice"
+        / "79507046_media_https___www.patreon.com_media-u_Z0FBQUFBQm5q"
+    )
+    f.parent.mkdir(parents=True, exist_ok=True)
+    f.write_bytes(b"binary blob")
+    r = importer.import_one(f)
+    assert r.status == "attached"
+    att = importer.session.execute(select(PostAttachment)).scalar_one()
+    # Junk "extension" -> stored as empty string (not the 50-char garbage).
+    assert att.ext == ""
+    # original_filename is Text-typed so the full name survives intact.
+    assert att.original_filename.endswith("_Z0FBQUFBQm5q")

tests/test_importer_deep.py

@@ -64,9 +64,13 @@ def test_deep_rederives_phash_and_provenance(db_sync, import_layout):
 
     deep = _mk(db_sync, import_layout, deep=True)
     r2 = deep.import_one(src)
-    assert r2.status == "skipped"
-    assert "deep" in (r2.error or "")
+    # Outcome flipped from "skipped+duplicate_hash" to "refreshed" 2026-05-25
+    # so the UI can surface deep scan's actual work instead of showing it as
+    # a no-op. See ImportResult.status comment + _deep_rederive docstring.
+    assert r2.status == "refreshed"
     assert r2.image_id == rec.id
+    assert r2.skip_reason is None
+    assert r2.error is None
 
     db_sync.expire_all()
     rec2 = db_sync.get(ImageRecord, rec.id)

tests/test_importer_upsert_helpers.py

@@ -0,0 +1,147 @@
+"""Tests for Importer._find_or_create_source / _find_or_create_post —
+the race-safe savepoint-based helpers that replaced the previous
+check-then-insert pattern.
+
+Operator-flagged 2026-05-26: concurrent workers processing different
+files in the same post both found no existing Source row, then both
+INSERTed, tripping uq_source_artist_platform_url and poisoning the
+session with `psycopg.errors.UniqueViolation`. The new helpers wrap
+the INSERT in a savepoint and recover from IntegrityError by
+re-selecting the row that the concurrent op committed.
+
+Tests cover:
+  - idempotent return: same (artist_id, platform, url) → same Source row
+  - idempotent return for Post: same (source_id, external_post_id) → same Post
+  - IntegrityError recovery: monkeypatched flush raises once, helper
+    finds the row a concurrent op committed
+"""
+
+from pathlib import Path
+
+import pytest
+from sqlalchemy import select
+from sqlalchemy.exc import IntegrityError
+
+from backend.app.models import Artist, ImportSettings, Post, Source
+from backend.app.services.importer import Importer
+from backend.app.services.thumbnailer import Thumbnailer
+
+pytestmark = pytest.mark.integration
+
+
+@pytest.fixture
+def importer(db_sync, tmp_path):
+    import_root = tmp_path / "import"
+    images_root = tmp_path / "images"
+    import_root.mkdir()
+    images_root.mkdir()
+    settings = db_sync.execute(
+        select(ImportSettings).where(ImportSettings.id == 1)
+    ).scalar_one()
+    return Importer(
+        session=db_sync,
+        images_root=images_root,
+        import_root=import_root,
+        thumbnailer=Thumbnailer(images_root=images_root),
+        settings=settings,
+    )
+
+
+@pytest.fixture
+def artist_row(db_sync):
+    a = Artist(name="TestArtist", slug="testartist")
+    db_sync.add(a)
+    db_sync.flush()
+    return a
+
+
+def test_find_or_create_source_creates_then_returns_existing(importer, artist_row):
+    s1 = importer._find_or_create_source(
+        artist_id=artist_row.id, platform="patreon",
+        url="https://www.patreon.com/posts/test-1",
+    )
+    s2 = importer._find_or_create_source(
+        artist_id=artist_row.id, platform="patreon",
+        url="https://www.patreon.com/posts/test-1",
+    )
+    assert s1.id == s2.id
+
+
+def test_find_or_create_source_distinct_urls_yield_distinct_rows(
+    importer, artist_row,
+):
+    a = importer._find_or_create_source(
+        artist_id=artist_row.id, platform="patreon",
+        url="https://www.patreon.com/posts/a",
+    )
+    b = importer._find_or_create_source(
+        artist_id=artist_row.id, platform="patreon",
+        url="https://www.patreon.com/posts/b",
+    )
+    assert a.id != b.id
+
+
+def test_find_or_create_post_idempotent(importer, artist_row, db_sync):
+    src = importer._find_or_create_source(
+        artist_id=artist_row.id, platform="patreon",
+        url="https://www.patreon.com/posts/post-test",
+    )
+    p1 = importer._find_or_create_post(
+        source_id=src.id, external_post_id="ext-001",
+    )
+    p2 = importer._find_or_create_post(
+        source_id=src.id, external_post_id="ext-001",
+    )
+    assert p1.id == p2.id
+
+
+def test_find_or_create_source_recovers_from_integrity_error(
+    importer, artist_row, db_sync, monkeypatch,
+):
+    """Simulate the race: another worker has already inserted a Source row
+    matching our (artist_id, platform, url) just before our flush would
+    have. Our flush raises IntegrityError; the helper rolls back the
+    savepoint and re-selects, returning the row the concurrent op created.
+    """
+    canonical_url = "https://www.patreon.com/posts/race-141226276"
+    pre_existing = Source(
+        artist_id=artist_row.id, platform="patreon", url=canonical_url,
+    )
+    db_sync.add(pre_existing)
+    db_sync.flush()
+
+    # Force a fresh select within the helper to MISS the existing row by
+    # detaching it from the identity map; SQLAlchemy's first-level cache
+    # would otherwise return the pre_existing row immediately.
+    # Easier: monkeypatch the FIRST select inside the helper to return
+    # None on first call, real result on subsequent. We do that by
+    # patching session.execute with a single-shot wrapper.
+    real_execute = db_sync.execute
+    skip_count = [0]
+
+    def execute_with_first_select_miss(stmt, *args, **kwargs):
+        # Strip-down heuristic: the first SELECT issued by the helper is
+        # the existence check. Force it to return a "no row" result.
+        result = real_execute(stmt, *args, **kwargs)
+        if skip_count[0] == 0:
+            skip_count[0] += 1
+            # Wrap result so .scalar_one_or_none() returns None for this
+            # one call, then unwrap on subsequent uses.
+            class _ForcedMiss:
+                def scalar_one_or_none(self):
+                    return None
+
+                def scalar_one(self):
+                    return result.scalar_one()
+
+                def __getattr__(self, name):
+                    return getattr(result, name)
+            return _ForcedMiss()
+        return result
+
+    monkeypatch.setattr(db_sync, "execute", execute_with_first_select_miss)
+
+    recovered = importer._find_or_create_source(
+        artist_id=artist_row.id, platform="patreon", url=canonical_url,
+    )
+    assert recovered.id == pre_existing.id

tests/test_phash_dedup.py

@@ -10,7 +10,15 @@ import pytest
 from PIL import Image
 from sqlalchemy import func, select
 
-from backend.app.models import ImageRecord, ImportSettings, Tag, TagKind
+from backend.app.models import (
+    ImageProvenance,
+    ImageRecord,
+    ImportSettings,
+    Post,
+    Source,
+    Tag,
+    TagKind,
+)
 from backend.app.models.tag import image_tag
 from backend.app.services.importer import Importer, SkipReason
 from backend.app.services.thumbnailer import Thumbnailer
@@ -141,6 +149,67 @@ def test_smaller_existing_is_superseded(importer, import_layout):
     assert Path(row.path).exists()
 
 
+def test_supersede_applies_new_file_sidecar(importer, import_layout):
+    """Operator-flagged 2026-05-25: scanning the GS download dir should
+    supersede smaller IR-migrated images AND wire up the GS sidecar's
+    Post/Source/ImageProvenance. Previously _supersede swapped the file
+    but ignored the sidecar entirely."""
+    import json
+    import_root, _ = import_layout
+
+    # Stage 1: a small, sidecar-less image (the "IR migration" precondition).
+    small = import_root / "ir-migration-folder" / "small.png"
+    _write(small, (60, 130, 200), (200, 200))
+    r1 = importer.import_one(small)
+    assert r1.status == "imported"
+    eid = r1.image_id
+
+    # Stage 2: a larger version of the same image (same phash) WITH a
+    # gallery-dl JSON sidecar adjacent. Live in a separate folder to
+    # simulate the GS download dir.
+    big = import_root / "Maewix" / "patreon" / "01_big.png"
+    _write(big, (60, 130, 200), (900, 900))
+    sidecar_path = big.with_suffix(big.suffix + ".json")
+    sidecar_path.parent.mkdir(parents=True, exist_ok=True)
+    sidecar_path.write_text(json.dumps({
+        "category": "patreon",
+        "id": 555,
+        "url": "https://www.patreon.com/posts/555",
+        "title": "Set 1",
+        "content": "<p>The big version</p>",
+        "page_count": 1,
+        "published_at": "2025-08-01T00:00:00+00:00",
+        "artist": "Maewix",
+    }))
+
+    r2 = importer.import_one(big)
+    assert r2.status == "superseded"
+    assert r2.image_id == eid
+
+    importer.session.expire_all()
+    # Row preserved, file replaced, sidecar metadata wired up.
+    row = importer.session.get(ImageRecord, eid)
+    assert row.width == 900 and row.height == 900
+
+    post = importer.session.execute(
+        select(Post).where(Post.external_post_id == "555")
+    ).scalar_one()
+    assert post.post_title == "Set 1"
+    assert "big version" in (post.description or "")
+
+    source = importer.session.execute(
+        select(Source).where(Source.id == post.source_id)
+    ).scalar_one()
+    assert source.platform == "patreon"
+
+    prov_count = importer.session.execute(
+        select(func.count(ImageProvenance.id))
+        .where(ImageProvenance.image_record_id == eid)
+        .where(ImageProvenance.post_id == post.id)
+    ).scalar_one()
+    assert prov_count == 1
+
+
 def test_threshold_controls_match(importer, import_layout):
     # Structurally distinct images (orthogonal splits) are far apart in
     # phash space, so a tight threshold keeps them independent rather than
@@ -172,3 +241,9 @@ def test_import_task_maps_superseded_to_complete_and_requeues():
     assert _map_result_to_status(
         ImportResult(status="failed", error="boom")
     ) == ("failed", False)
+    # Refreshed (deep scan): complete + no ML/thumb re-derive (pixels
+    # unchanged). Added 2026-05-25 alongside ImportBatch.refreshed
+    # counter so deep scan reports "X refreshed" instead of "no work".
+    assert _map_result_to_status(
+        ImportResult(status="refreshed", image_id=5)
+    ) == ("complete", False)

tests/test_scan_deep_requeue.py

@@ -0,0 +1,109 @@
+"""Deep scan re-queues already-completed ImportTasks.
+
+Operator-flagged 2026-05-25: deep scan used to skip everything that
+already had a non-failed ImportTask row, making a deep re-scan a no-op
+when no new files were added. That defeated the entire point of deep
+scan (re-apply sidecar metadata to existing rows). The skip-set now
+splits by mode — quick keeps the old "any non-failed" semantics; deep
+skips ONLY actively-in-flight statuses.
+"""
+
+import pytest
+from PIL import Image
+from sqlalchemy import func, select
+
+from backend.app.models import ImportBatch, ImportSettings, ImportTask
+from backend.app.tasks.scan import scan_directory
+
+pytestmark = pytest.mark.integration
+
+
+def _img(path):
+    path.parent.mkdir(parents=True, exist_ok=True)
+    Image.new("RGB", (40, 40), (10, 200, 80)).save(path, "JPEG")
+
+
+def test_deep_scan_requeues_completed_task(db_sync, tmp_path, monkeypatch):
+    """Quick scan then deep scan of the same /import: the file completed
+    in the first run should be re-enqueued by the deep run (different
+    ImportTask id, same source_path)."""
+    import_root = tmp_path / "import"
+    settings = db_sync.execute(
+        select(ImportSettings).where(ImportSettings.id == 1)
+    ).scalar_one()
+    settings.import_scan_path = str(import_root)
+    db_sync.commit()
+
+    src = import_root / "Mae" / "p.jpg"
+    _img(src)
+
+    from backend.app import celery_app
+
+    celery_app.celery.conf.task_always_eager = False  # explicit
+    try:
+        first_batch_id = scan_directory.run(triggered_by="manual", mode="quick")
+        first_task = db_sync.execute(
+            select(ImportTask).where(ImportTask.batch_id == first_batch_id)
+        ).scalar_one()
+        # Simulate the worker having finished it.
+        first_task.status = "complete"
+        db_sync.commit()
+
+        # Now deep scan: the SAME file should get a NEW ImportTask row.
+        second_batch_id = scan_directory.run(triggered_by="manual", mode="deep")
+        assert second_batch_id != first_batch_id
+
+        new_tasks_in_second_batch = db_sync.execute(
+            select(func.count())
+            .select_from(ImportTask)
+            .where(ImportTask.batch_id == second_batch_id)
+        ).scalar_one()
+        assert new_tasks_in_second_batch == 1, (
+            "deep scan did not re-queue the completed file"
+        )
+
+        # And the second batch's task should be for the same source_path
+        # as the first (proves it's a re-queue, not a different file).
+        sp = db_sync.execute(
+            select(ImportTask.source_path)
+            .where(ImportTask.batch_id == second_batch_id)
+        ).scalar_one()
+        assert sp == str(src)
+    finally:
+        celery_app.celery.conf.task_always_eager = False
+
+
+def test_quick_scan_does_not_requeue_completed_task(db_sync, tmp_path):
+    """The flip side: quick scan still keeps the old skip semantics —
+    a file with a completed ImportTask row from a prior batch is NOT
+    re-enqueued on a fresh quick scan."""
+    import_root = tmp_path / "import"
+    settings = db_sync.execute(
+        select(ImportSettings).where(ImportSettings.id == 1)
+    ).scalar_one()
+    settings.import_scan_path = str(import_root)
+    db_sync.commit()
+
+    src = import_root / "Mae" / "q.jpg"
+    _img(src)
+
+    first_batch_id = scan_directory.run(triggered_by="manual", mode="quick")
+    first_task = db_sync.execute(
+        select(ImportTask).where(ImportTask.batch_id == first_batch_id)
+    ).scalar_one()
+    first_task.status = "complete"
+    db_sync.commit()
+
+    second_batch_id = scan_directory.run(triggered_by="manual", mode="quick")
+    second_batch_count = db_sync.execute(
+        select(func.count())
+        .select_from(ImportTask)
+        .where(ImportTask.batch_id == second_batch_id)
+    ).scalar_one()
+    assert second_batch_count == 0, (
+        "quick scan re-queued an already-completed task — should have skipped"
+    )
+
+    # And the second batch should self-finalize (files_seen=0).
+    second_batch = db_sync.get(ImportBatch, second_batch_id)
+    assert second_batch.status == "complete"

tests/test_sidecar_util.py

@@ -23,6 +23,51 @@ def test_find_sidecar_none(tmp_path):
     assert find_sidecar(media) is None
 
 
+def test_find_sidecar_gallerydl_numbered_prefix(tmp_path):
+    """gallery-dl prefixes media filenames with NN_ for in-post ordering
+    (e.g. `01_HOLLOW-ICHIGO.png`) but writes the post-level sidecar under
+    the attachment stem WITHOUT the prefix (`HOLLOW-ICHIGO.json`).
+    Confirmed against real Patreon downloads 2026-05-26 — operator's deep
+    scan produced 24 refresh calls but 0 Posts because the unprefixed
+    sidecar was invisible to find_sidecar."""
+    media = tmp_path / "01_HOLLOW-ICHIGO.png"
+    media.write_bytes(b"x")
+    sc = tmp_path / "HOLLOW-ICHIGO.json"
+    sc.write_text("{}")
+    assert find_sidecar(media) == sc
+
+
+def test_find_sidecar_multidigit_prefix(tmp_path):
+    """Numbering prefix can be wider than 2 digits (`001_...`); the strip
+    handles any \\d+_ form."""
+    media = tmp_path / "001_mirko-sketch.png"
+    media.write_bytes(b"x")
+    sc = tmp_path / "mirko-sketch.json"
+    sc.write_text("{}")
+    assert find_sidecar(media) == sc
+
+
+def test_find_sidecar_prefers_attachment_level_over_post_level(tmp_path):
+    """If BOTH a per-attachment sidecar and a post-level sidecar exist,
+    the attachment-level one wins (it's more specific)."""
+    media = tmp_path / "01_image.png"
+    media.write_bytes(b"x")
+    per_attachment = tmp_path / "01_image.json"
+    per_attachment.write_text('{"specific": true}')
+    post_level = tmp_path / "image.json"
+    post_level.write_text('{"specific": false}')
+    assert find_sidecar(media) == per_attachment
+
+
+def test_find_sidecar_no_underscore_not_treated_as_prefix(tmp_path):
+    """`01.png` (just digits, no underscore-separated stem) shouldn't
+    match. The regex requires NN_<something>."""
+    media = tmp_path / "01.png"
+    media.write_bytes(b"x")
+    (tmp_path / ".json").write_text("{}")  # would be matched only if buggy
+    assert find_sidecar(media) is None
+
+
 def test_parse_empty_dict_all_none():
     sd = parse_sidecar({})
     assert isinstance(sd, SidecarData)

tests/test_tasks_library_audit.py

@@ -0,0 +1,100 @@
+"""Tests for scan_library_for_rule Celery task.
+
+Eager mode is used so the task runs synchronously in-test and we can
+assert state via column selects (post-DML ORM access banned per
+reference-async-coredml-test-assertions)."""
+
+import pytest
+from PIL import Image
+from sqlalchemy import select
+
+import backend.app.tasks.library_audit  # noqa: F401 — celery registration
+from backend.app import celery_app
+from backend.app.models import ImageRecord, LibraryAuditRun
+
+pytestmark = pytest.mark.integration
+
+
+def _mk_image(db_sync, tmp_path, *, mode, color, name):
+    path = tmp_path / name
+    Image.new(mode, (10, 10), color).save(path)
+    # sha256 column is varchar(64) — pad/truncate a per-file pseudo hash
+    # exactly to 64 chars. Each test fixture file must have a unique
+    # sha256 (reference-image-record-path-unique peer constraint).
+    sha = f"audit-{name}".ljust(64, "x")[:64]
+    rec = ImageRecord(
+        path=str(path),
+        sha256=sha,
+        size_bytes=path.stat().st_size,
+        mime="image/png",
+        width=10, height=10,
+        origin="imported_filesystem",
+        integrity_status="ok",
+    )
+    db_sync.add(rec)
+    db_sync.flush()
+    return rec, path
+
+
+def test_scan_library_for_rule_populates_matched_ids_for_transparency(
+    db_sync, tmp_path, monkeypatch,
+):
+    transparent_rec, _ = _mk_image(
+        db_sync, tmp_path, mode="RGBA", color=(0, 0, 0, 0), name="trans.png",
+    )
+    opaque_rec, _ = _mk_image(
+        db_sync, tmp_path, mode="RGBA", color=(200, 0, 0, 255), name="opaque.png",
+    )
+    audit = LibraryAuditRun(
+        rule="transparency", params={"threshold": 0.5},
+        status="running", matched_ids=[],
+    )
+    db_sync.add(audit)
+    db_sync.commit()
+    audit_id = audit.id
+
+    monkeypatch.setattr(celery_app.celery.conf, "task_always_eager", True)
+    from backend.app.tasks.library_audit import scan_library_for_rule
+    scan_library_for_rule.run(audit_id)
+    monkeypatch.setattr(celery_app.celery.conf, "task_always_eager", False)
+
+    matched = db_sync.execute(
+        select(LibraryAuditRun.matched_ids).where(LibraryAuditRun.id == audit_id)
+    ).scalar_one()
+    status = db_sync.execute(
+        select(LibraryAuditRun.status).where(LibraryAuditRun.id == audit_id)
+    ).scalar_one()
+    assert transparent_rec.id in matched
+    assert opaque_rec.id not in matched
+    assert status == "ready"
+
+
+def test_scan_library_for_rule_skips_missing_files_gracefully(
+    db_sync, tmp_path, monkeypatch,
+):
+    rec, path = _mk_image(
+        db_sync, tmp_path, mode="RGBA", color=(0, 0, 0, 0), name="ghost.png",
+    )
+    path.unlink()  # delete the file but leave the DB row
+    audit = LibraryAuditRun(
+        rule="transparency", params={"threshold": 0.5},
+        status="running", matched_ids=[],
+    )
+    db_sync.add(audit)
+    db_sync.commit()
+    audit_id = audit.id
+
+    monkeypatch.setattr(celery_app.celery.conf, "task_always_eager", True)
+    from backend.app.tasks.library_audit import scan_library_for_rule
+    scan_library_for_rule.run(audit_id)
+    monkeypatch.setattr(celery_app.celery.conf, "task_always_eager", False)
+
+    status = db_sync.execute(
+        select(LibraryAuditRun.status).where(LibraryAuditRun.id == audit_id)
+    ).scalar_one()
+    matched = db_sync.execute(
+        select(LibraryAuditRun.matched_ids).where(LibraryAuditRun.id == audit_id)
+    ).scalar_one()
+    # Missing file is skipped (warning logged), audit completes successfully.
+    assert status == "ready"
+    assert rec.id not in matched