feat: pzmm conflict detection + content-type categorization

- mod_files manifest table populated at parse time
- POST /api/conflicts endpoint
- mod_types fingerprinting feeds derive_category
- DD filelist regex broadened to cover conflict-eligible exts
- media/maps/<*>/* excluded from manifest (per-mod namespaced,
  no conflict value, can be tens of MB per mod)

Plan: docs/plans/2026-05-04-pzmm-conflict-and-typing.md

.gitignore

@@ -20,6 +20,9 @@ __pycache__/
 *.bak
 *.bak-*
 
+# Operational DB dumps (pg_dump output kept locally for rollback)
+backups/
+
 # Editor / IDE / OS artifacts
 *.swp
 *.swo

api/app.py

@@ -26,6 +26,7 @@ from pydantic import BaseModel, Field
 
 import adapters
 import db
+import diagnostics
 import expansion
 import jobs
 import steam
@@ -191,6 +192,7 @@ def _row_to_modinfo(r) -> ModInfo:
         maps=list(r["maps"] or []),
         is_addon=bool(r["is_addon"]) if "is_addon" in r else False,
         workshop_tags=list(r["workshop_tags"] or []) if "workshop_tags" in r else [],
+        mod_types=list(r["mod_types"] or []) if "mod_types" in r else [],
     )
 
 
@@ -684,7 +686,7 @@ async def _build_result_for_job(
         SELECT mp.workshop_id, mp.mod_id, mp.name, mp.category,
                mp.requirements, mp.load_after, mp.load_before,
                mp.incompatible_mods, mp.load_first, mp.load_last,
-               mp.tags, mp.maps, mp.is_addon, wm.tags AS workshop_tags
+               mp.tags, mp.maps, mp.is_addon, mp.mod_types, wm.tags AS workshop_tags
           FROM mod_parsed mp
           JOIN workshop_meta wm ON wm.workshop_id = mp.workshop_id
          WHERE mp.workshop_id = ANY($1::text[])
@@ -1001,7 +1003,7 @@ async def sort_endpoint(req: SortRequest, request: Request) -> Dict[str, Any]:
                 SELECT mp.workshop_id, mp.mod_id, mp.name, mp.category,
                        mp.requirements, mp.load_after, mp.load_before,
                        mp.incompatible_mods, mp.load_first, mp.load_last,
-                       mp.tags, mp.maps, mp.is_addon, wm.tags AS workshop_tags
+                       mp.tags, mp.maps, mp.is_addon, mp.mod_types, wm.tags AS workshop_tags
                 FROM mod_parsed mp
                 JOIN workshop_meta wm ON wm.workshop_id = mp.workshop_id
                 WHERE mp.workshop_id = ANY($1::text[])
@@ -1215,7 +1217,7 @@ async def resort_endpoint(req: ResortRequest, request: Request) -> Dict[str, Any
             SELECT mp.workshop_id, mp.mod_id, mp.name, mp.category,
                    mp.requirements, mp.load_after, mp.load_before,
                    mp.incompatible_mods, mp.load_first, mp.load_last,
-                   mp.tags, mp.maps, mp.is_addon, wm.tags AS workshop_tags
+                   mp.tags, mp.maps, mp.is_addon, mp.mod_types, wm.tags AS workshop_tags
             FROM mod_parsed mp
             JOIN workshop_meta wm ON wm.workshop_id = mp.workshop_id
             WHERE mp.workshop_id IN (SELECT workshop_id FROM selected_wsids)
@@ -1469,6 +1471,71 @@ async def vote_broken_mod(
     return {"upvotes": int(row["upvotes"]), "downvotes": int(row["downvotes"])}
 
 
+@app.post("/api/conflicts")
+async def conflicts_endpoint(req: SortRequest, request: Request) -> Dict[str, Any]:
+    """Detect rel_paths claimed by ≥2 input mods with non-equal sha1.
+
+    v1: bare wsids only. Collection input returns 400 so the caller can
+    resolve via /api/sort first (where the async-job + drain-progress
+    plumbing already lives). Mods whose `files_manifest_built` is false
+    cannot be analyzed and are reported in `missing_manifests` instead of
+    silently ignored.
+    """
+    bare_wsids, collection_ids = parse_with_collections(req.input or "")
+    if collection_ids:
+        raise HTTPException(
+            status_code=400,
+            detail="conflict scan does not support collection input; resolve via /api/sort first",
+        )
+    if not bare_wsids:
+        raise HTTPException(status_code=400, detail="no workshop ids found in input")
+    if len(bare_wsids) > MAX_IDS:
+        raise HTTPException(
+            status_code=413,
+            detail=f"too many workshop ids ({len(bare_wsids)} > {MAX_IDS})",
+        )
+
+    pool = request.app.state.db
+    async with pool.acquire() as conn:
+        rows = await conn.fetch(
+            """
+            SELECT mp.workshop_id, mp.mod_id, mp.name, mp.category,
+                   mp.requirements, mp.load_after, mp.load_before,
+                   mp.incompatible_mods, mp.load_first, mp.load_last,
+                   mp.tags, mp.maps, mp.is_addon, mp.mod_types,
+                   mp.files_manifest_built, wm.tags AS workshop_tags
+              FROM mod_parsed mp
+              JOIN workshop_meta wm ON wm.workshop_id = mp.workshop_id
+             WHERE mp.workshop_id = ANY($1::text[])
+               AND mp.parsed_at_time_updated = wm.time_updated
+             ORDER BY mp.workshop_id, mp.mod_id
+            """,
+            bare_wsids,
+        )
+
+        mods: List[ModInfo] = [_row_to_modinfo(r) for r in rows]
+
+        # Missing-manifest wsids: input wsids that have no mod_parsed rows
+        # OR whose rows all have files_manifest_built=false. Any single
+        # built row in a multi-mod wsid counts as "manifest available".
+        wsid_has_manifest: Dict[str, bool] = {}
+        for r in rows:
+            w = r["workshop_id"]
+            built = bool(r["files_manifest_built"])
+            wsid_has_manifest[w] = wsid_has_manifest.get(w, False) or built
+        missing_manifests = [w for w in bare_wsids if not wsid_has_manifest.get(w, False)]
+
+        conflicts = await diagnostics.scan_file_conflicts(conn, mods)
+
+    return {
+        "conflicts": [
+            {"rel_path": c.rel_path, "providers": c.providers, "winner": c.winner}
+            for c in conflicts
+        ],
+        "missing_manifests": missing_manifests,
+    }
+
+
 # ── static frontend ────────────────────────────────────────────────────────
 # Mount LAST so all API routes win path resolution.
 _FRONTEND_DIR = Path(__file__).resolve().parent.parent / "frontend"

api/categorize.py

@@ -0,0 +1,58 @@
+"""Public helper for mapping pzmm content-type tags to sortof CATEGORY_ORDER.
+
+The same mapping is also inlined in `mlos_sort.py` (both api/ and worker/
+copies, deliberately — worker uses a separate venv with no FastAPI deps,
+so it cannot import from api/). This module exposes the helper for
+non-mlos consumers (e.g. /api/conflicts diagnostics output) without
+forcing them to drag in the whole sorter module.
+
+Source: pzmm core/mods.py:detect_mod_types ordering, mapped to sortof's
+CATEGORY_ORDER buckets per docs/plans/2026-05-04-pzmm-conflict-and-typing.md
+§3.4.
+"""
+
+from __future__ import annotations
+
+from typing import Dict, List, Optional
+
+# Items / Animations / Lua / Unknown intentionally absent — too generic to
+# drive a category decision; callers should fall through to other heuristics.
+_TYPE_TO_CAT: Dict[str, str] = {
+    "Maps":         "map",
+    "Vehicles":     "vehicle",
+    "Weapons":      "weapon",
+    "Clothing":     "wearable",
+    "Traits":       "code",
+    "Professions":  "profession",
+    "Recipes":      "crafting",
+    "Tiles":        "tile",
+    "Textures":     "texture",
+    "Sounds":       "sound",
+    "UI":           "ui",
+    "Translations": "translation",
+    "Patch":        "patch",
+    "Dependency":   "tweaks",
+    "Framework":    "tweaks",
+}
+
+
+def types_to_category(mod_types: List[str], name: str = "") -> Optional[str]:
+    """First mod_type that maps to a sortof CATEGORY_ORDER bucket wins.
+
+    Returns the bucket name (e.g. "weapon", "vehicle"), or None when:
+      - mod_types is empty (manifest not yet built), or
+      - mod_types contains only skip-types (Items / Animations / Lua / Unknown).
+
+    The `name` arg is used for the vehicle_spawn refinement only — when a
+    Vehicles-tagged mod is named like "spawn zone X", the more specific
+    `vehicle_spawn` bucket wins over the generic `vehicle`.
+    """
+    if not mod_types:
+        return None
+    for t in mod_types:
+        cat = _TYPE_TO_CAT.get(t)
+        if cat:
+            if cat == "vehicle" and name and "spawn zone" in name.lower():
+                return "vehicle_spawn"
+            return cat
+    return None

api/diagnostics.py

@@ -0,0 +1,93 @@
+"""File-level conflict detection from cached manifests.
+
+Port of pzmm core/scanner.py:scan_file_conflicts adapted to read from the
+mod_files table (populated by worker.build_manifest_and_types) instead of
+walking on-disk media trees. See docs/plans/2026-05-04-pzmm-conflict-and-typing.md.
+"""
+
+from __future__ import annotations
+
+from collections import defaultdict
+from dataclasses import dataclass
+from typing import Dict, List, Tuple
+
+from mlos_sort import ModInfo
+
+
+@dataclass
+class FileConflict:
+    rel_path: str
+    providers: List[str]   # mod_ids in input load order
+    winner: str            # mod_id (last in load order)
+
+
+_FETCH_MANIFEST = """
+WITH inputs AS (
+    SELECT unnest($1::text[]) AS workshop_id,
+           unnest($2::text[]) AS mod_id
+)
+SELECT mf.workshop_id, mf.mod_id, mf.rel_path, mf.sha1
+  FROM mod_files mf
+  JOIN inputs i
+    ON mf.workshop_id = i.workshop_id
+   AND mf.mod_id      = i.mod_id
+"""
+
+
+async def scan_file_conflicts(conn, mods: List[ModInfo]) -> List[FileConflict]:
+    """For the given (already-loaded) ModInfos, report rel_paths claimed by
+    ≥2 mods with non-equal sha1. Returns list ordered by rel_path.
+
+    Mods without manifest rows (`files_manifest_built=false`) silently
+    contribute nothing to the conflict scan; the caller is responsible for
+    surfacing them as `missing_manifests` in any user-facing payload.
+    """
+    if len(mods) < 2:
+        return []
+
+    wsids = [m.workshop_id or "" for m in mods]
+    mod_ids = [m.id for m in mods]
+    rows = await conn.fetch(_FETCH_MANIFEST, wsids, mod_ids)
+
+    # mod_id → load-order index (input order = load order, mirroring pzmm)
+    order_index: Dict[str, int] = {m.id: i for i, m in enumerate(mods)}
+
+    # rel_path → list of (load_order_index, mod_id, sha1)
+    by_path: Dict[str, List[Tuple[int, str, str]]] = defaultdict(list)
+    for r in rows:
+        mod_id = r["mod_id"]
+        idx = order_index.get(mod_id)
+        if idx is None:
+            continue
+        by_path[r["rel_path"]].append((idx, mod_id, r["sha1"]))
+
+    conflicts: List[FileConflict] = []
+    for rel, entries in by_path.items():
+        # Need ≥2 distinct providers AND ≥2 distinct sha1s. If every
+        # provider ships byte-identical content (same sha1), it's a
+        # duplicate, not a conflict — pzmm scanner.py:55–66.
+        unique_providers = {mod_id for _, mod_id, _ in entries}
+        if len(unique_providers) < 2:
+            continue
+        unique_hashes = {sha for _, _, sha in entries}
+        if len(unique_hashes) < 2:
+            continue
+        # Order providers by input load-order index. Winner = last loaded.
+        ordered = sorted(entries, key=lambda e: e[0])
+        providers = [mod_id for _, mod_id, _ in ordered]
+        # De-dup providers preserving order (a mod could ship the same
+        # rel_path under both B41 and B42 layouts → seen twice).
+        seen: set = set()
+        dedup_providers: List[str] = []
+        for p in providers:
+            if p not in seen:
+                seen.add(p)
+                dedup_providers.append(p)
+        conflicts.append(FileConflict(
+            rel_path=rel,
+            providers=dedup_providers,
+            winner=dedup_providers[-1],
+        ))
+
+    conflicts.sort(key=lambda c: c.rel_path)
+    return conflicts

api/mlos_sort.py

@@ -130,6 +130,11 @@ class ModInfo:
     # signal for build / multiplayer / category detection. Distinct from
     # `tags` which is mod.info-side (freeform).
     workshop_tags: List[str] = field(default_factory=list)
+    # pzmm-style content fingerprint (Maps, Vehicles, Weapons, Traits, …)
+    # populated by worker.build_manifest_and_types at parse time. Empty when
+    # files_manifest_built=false (older cached rows); derive_category falls
+    # through to the existing cascade in that case.
+    mod_types: List[str] = field(default_factory=list)
     warnings: Dict[str, List[str]] = field(default_factory=dict)
 
 
@@ -389,30 +394,77 @@ def _name_has(name: str, hints: List[str]) -> bool:
     return any(h in n for h in hints)
 
 
+# pzmm content-type → sortof CATEGORY_ORDER mapping. "skip" entries fall
+# through to the existing derive_category cascade. Items/Animations/Lua/Unknown
+# are too generic; Maps/Sounds/Patch/Vehicles/Clothing duplicate signals already
+# captured by the cascade but stay here as fallbacks for poorly-tagged mods.
+_TYPE_TO_CAT: Dict[str, str] = {
+    "Maps":         "map",
+    "Vehicles":     "vehicle",
+    "Weapons":      "weapon",
+    "Clothing":     "wearable",
+    "Traits":       "code",
+    "Professions":  "profession",
+    "Recipes":      "crafting",
+    "Tiles":        "tile",
+    "Textures":     "texture",
+    "Sounds":       "sound",
+    "UI":           "ui",
+    "Translations": "translation",
+    "Patch":        "patch",
+    "Dependency":   "tweaks",
+    "Framework":    "tweaks",
+}
+
+
+def _types_to_category(mod_types: List[str], name: str) -> Optional[str]:
+    """First mod_type that maps to a sortof CATEGORY_ORDER bucket wins.
+    Returns None if mod_types contains only skip-types (Items/Animations/Lua/
+    Unknown), so the caller can fall through to the existing cascade."""
+    for t in mod_types:
+        cat = _TYPE_TO_CAT.get(t)
+        if cat:
+            # vehicle_spawn refinement matches the downstream ws_tag check.
+            if cat == "vehicle" and name and "spawn zone" in name.lower():
+                return "vehicle_spawn"
+            return cat
+    return None
+
+
 def derive_category(mod: ModInfo) -> str:
     """Best-effort category from mod.info + workshop_meta.tags + name.
 
     Detection order (most specific → least):
       1. mod.info `category=` if explicit and recognized.
-      2. patch / fix name regex (Spec G-patch).
-      3. library/framework name regex (extends FRAMEWORK_KEYS).
-      4. mod.maps non-empty → map.
-      5. moodle / profession / movement / specific gameplay axes by name.
-      6. Workshop tags (canonical Steam controlled vocab): Audio + 'music' →
+      2. pzmm-style mod_types fingerprint (when files_manifest_built=true).
+      3. patch / fix name regex (Spec G-patch).
+      4. library/framework name regex (extends FRAMEWORK_KEYS).
+      5. mod.maps non-empty → map.
+      6. moodle / profession / movement / specific gameplay axes by name.
+      7. Workshop tags (canonical Steam controlled vocab): Audio + 'music' →
          music; Audio → sound; Weapons → weapon; Vehicles → vehicle;
          Clothing/Armor + 'armor' → armor, else wearable; Building →
          building; Farming → farming; Food → food; Skills → profession
          (or moodle); Interface → ui; Textures → texture;
          Language/Translation → translation; QOL → qol; Multiplayer alone
          → multiplayer.
-      7. mod.info tags (freeform fallback).
-      8. FRAMEWORK_KEYS substring match → tweaks.
-      9. Default → other.
+      8. mod.info tags (freeform fallback).
+      9. FRAMEWORK_KEYS substring match → tweaks.
+     10. Default → other.
     """
     if mod.category in CATEGORY_ORDER and mod.category != "undefined":
         return mod.category
 
     name = mod.name or ""
+
+    # pzmm-style content fingerprint takes precedence over name regex when
+    # available. Empty mod_types means files_manifest_built=false (older
+    # cached row); fall through to existing cascade.
+    if mod.mod_types:
+        cat = _types_to_category(mod.mod_types, name)
+        if cat:
+            return cat
+
     if name and _PATCH_NAME_RE.search(name):
         return "patch"
     if _name_has(name, _LIB_NAME_HINTS) or (name and _LIB_NAME_RE.search(name)):

docs/plans/2026-05-04-pzmm-conflict-and-typing.md

@@ -0,0 +1,343 @@
+# Plan: pzmm conflict detection + content-type categorization
+
+**Date:** 2026-05-04
+**Branch:** `feat/pzmm-conflict-typing`
+**Status:** Approved (Sam, 2026-05-04)
+
+**Sources read:**
+- `/tmp/pzmm-src/pzmm-main/core/scanner.py` — `scan_file_conflicts`, `solve_load_order`, `FileConflict`
+- `/tmp/pzmm-src/pzmm-main/core/mods.py` — `detect_mod_types`, `ModInfo`
+- `/tmp/pzmm-src/pzmm-main/core/bundle.py` — debug bundle (read for context, not integrated)
+- `/opt/sortof/init/01_schema.sql` and migrations 02..08
+- `/opt/sortof/api/app.py` — `/api/sort`, `_build_result_for_job`, `_row_to_modinfo`
+- `/opt/sortof/api/mlos_sort.py` — `CATEGORY_ORDER`, `derive_category`
+- `/opt/sortof/api/adapters.py` — `CAT_MAP`
+- `/opt/sortof/worker/worker.py` — `process_one`
+
+**Open questions resolved at approval:**
+- Manifest scope: walk all `media/` subtrees under the mod_id root, last-wins on duplicate rel_paths, **no per-branch column**.
+- `mod_files.size_bytes` column: keep.
+- Module split: `api/diagnostics.py` and `api/categorize.py` are **separate files**.
+- `/api/conflicts` v1: **bare wsids only**, return HTTP 400 on collection input. Defer async-job/collection-expansion plumbing to a follow-up plan.
+
+---
+
+## 1. Context
+
+pzmm ships two pieces sortof doesn't have today:
+
+1. **File-conflict detection** — when two mods both ship `media/scripts/items_food.txt` with byte-different content, the later one silently overrides the earlier one at runtime. PZ never reports this; the player only sees the symptom (broken food, duplicate item ids, etc.). pzmm walks each mod's `media/` tree, hashes the conflict-prone extensions (`.lua`, `.txt`, `.xml`, `.json`, `.ini`), and reports rel-paths claimed by ≥2 mods with non-equal content. Sortof currently only detects `mod_id` collisions (one mod_id under multiple wsids). File-level overrides are invisible to us.
+2. **Content-type detection** — pzmm walks `media/` paths plus the contents of `lua/` and `scripts/*.txt|xml` files to fingerprint what a mod actually ships (Weapons, Vehicles, Maps, Traits, Professions, Recipes, etc.). Sortof's `derive_category` infers category from `workshop_meta.tags` + name regex + `mod.info` tags. Authors who tag poorly (or skip tagging) end up in `other`/`undefined`. Detection from media/ contents is more reliable for those.
+
+Both pzmm functions assume on-disk media trees. Sortof's worker uses `tempfile.TemporaryDirectory` (`worker/worker.py:472`) — the entire DD extraction is destroyed at the end of `process_one`'s `with` block. **Only `mod.info` (as `raw_mod_info`), discovered map folder names, and a few derived columns persist.**
+
+This plan keeps the existing model: parse once, serve from DB. We **persist a manifest at parse time**. Re-fetch on demand was rejected — every conflict check would queue N DD pulls, minutes per request, completely unusable.
+
+We **do not import pzmm's `solve_load_order`**. Sortof's `mlos_sort.py` is strictly more correct (preorder, loadFirst/loadLast tiers, category buckets, patch G-axis, multi-branch picker, addon injection). pzmm's solver is a plain Kahn topo sort with no tie-breakers.
+
+---
+
+## 2. Integration A — File conflict detection
+
+### 2.1 New schema (`init/09_mod_files.sql`)
+
+```sql
+CREATE TABLE IF NOT EXISTS mod_files (
+    workshop_id  TEXT NOT NULL,
+    mod_id       TEXT NOT NULL,
+    rel_path     TEXT NOT NULL,            -- lowercased, posix-style, relative to mod_id root
+    sha1         TEXT NOT NULL,
+    size_bytes   INTEGER NOT NULL DEFAULT 0,
+    PRIMARY KEY (workshop_id, mod_id, rel_path),
+    FOREIGN KEY (workshop_id, mod_id) REFERENCES mod_parsed (workshop_id, mod_id) ON DELETE CASCADE
+);
+CREATE INDEX IF NOT EXISTS mod_files_rel_path_idx ON mod_files (rel_path);
+CREATE INDEX IF NOT EXISTS mod_files_mod_idx ON mod_files (workshop_id, mod_id);
+```
+
+Plus additions to `mod_parsed`:
+
+```sql
+ALTER TABLE mod_parsed
+  ADD COLUMN IF NOT EXISTS mod_types TEXT[] NOT NULL DEFAULT '{}',
+  ADD COLUMN IF NOT EXISTS files_manifest_built BOOLEAN NOT NULL DEFAULT FALSE;
+```
+
+The flag lets `derive_category` and `/api/conflicts` know whether a mod has a manifest yet (graceful degradation while the cache backfills organically).
+
+### 2.2 Worker changes (`worker/worker.py`)
+
+In `process_one`, **inside the existing `with tempfile.TemporaryDirectory` block** (after `discover_mod_infos`, before the `with` exits):
+
+**Single-pass requirement:** the manifest build (Integration A) and `detect_mod_types` content sniffing (Integration B) **share one pass over the tempdir**. No two-pass implementations. The walk reads each file's bytes once: hash → manifest insert; concurrently inspect path + content for type signals. The output is the `mod_files` rows for that mod_id and the ordered `mod_types` list, both committed in the same transaction as the existing `UPSERT_MOD_PARSED`.
+
+For each `(workshop_id, mod_id)` pair we just upserted:
+
+1. Compute `mod_id_root`: the directory whose name equals `mod.id`. For B41 (`mods/<modId>/mod.info`) that's `mip.parent`; for B42 (`mods/<modId>/<branch>/mod.info`) that's `mip.parent.parent`. Detect via `mip.parent.name == mod.id`.
+2. Single recursive walk under `mod_id_root` covering every `media/` subtree (handles B42 `<branch>/media/` + `common/media/` together). For each file:
+   - If suffix matches `_CONFLICT_EXTS = {".lua", ".txt", ".xml", ".json", ".ini"}` (verbatim from pzmm `scanner.py:21`), compute sha1 (chunked reader, mirrors pzmm `_sha1`) and accumulate `(rel_path, sha1, size_bytes)`. **Last-wins** on duplicate rel_paths across branches.
+   - Concurrently, in the same loop, accumulate the path-based signals from pzmm `mods.py:detect_mod_types` (lines 88–115): `Maps`, `Tiles`, `Textures`, `Vehicles`, `Clothing`, `Sounds`, `UI`, `Animations`, `Translations`, `Lua`, plus collected `lua_text_parts` and `script_text_parts` blobs (capped at 60 lua × 64 KB and 80 script × 96 KB per pzmm).
+3. After the walk, run pzmm's content-blob checks (lines 117–136): weapon/vehicle/item/recipe/clothing/trait/profession signals from concatenated blobs. Resolve to `mod_types` ordered list (lines 138–145).
+4. DELETE existing `mod_files` rows for `(workshop_id, mod_id)` then bulk INSERT new rows.
+5. UPSERT `mod_parsed.mod_types` and set `files_manifest_built = true` for the row.
+
+The whole step adds disk-walk + hashing of small text files only — typical mod has 20–200 files in scope, hashing is cheap (≤100 KB each, sha1 ≈ 500 MB/s). Estimated cost: <500 ms per mod, well under the DD pull cost we're already paying.
+
+### 2.3 New module: `api/diagnostics.py`
+
+Port of pzmm `scan_file_conflicts` adapted to read from `mod_files` instead of walking disk:
+
+```python
+async def scan_file_conflicts(conn, mods: list[ModInfo]) -> list[FileConflict]:
+    """For the given (already-loaded) ModInfos, report rel_paths claimed
+    by ≥2 mods with non-equal sha1. Returns list ordered by rel_path."""
+```
+
+Implementation:
+1. `SELECT workshop_id, mod_id, rel_path, sha1 FROM mod_files WHERE (workshop_id, mod_id) IN (...)`.
+2. Group rows in Python by `rel_path`.
+3. For each group with ≥2 distinct mods, count distinct sha1s. If >1, emit a `FileConflict`.
+4. Winner = last in input order (mirrors pzmm's "last in load order wins").
+
+Dataclass:
+```python
+@dataclass
+class FileConflict:
+    rel_path: str
+    providers: list[str]   # mod_ids (not ModInfo, to keep payload small)
+    winner: str            # mod_id
+```
+
+`pzmm.scanner._CONFLICT_EXTS` filtering happened at manifest-build time, so this read path doesn't need it.
+
+### 2.4 New endpoint: `POST /api/conflicts`
+
+Same input shape as `/api/sort`, **bare wsids only** (Q4 resolved):
+```json
+{"input": "wsid1;wsid2;wsid3", "rules": "...", "pz_build": "B42"}
+```
+
+If `parse_with_collections` returns any `collection_ids`, return HTTP 400 with `detail="conflict scan does not support collection input; resolve via /api/sort first"`.
+
+Response:
+```json
+{
+  "conflicts": [
+    {"rel_path": "media/scripts/items_food.txt",
+     "providers": ["FoodModA", "FoodModB"],
+     "winner": "FoodModB"}
+  ],
+  "missing_manifests": ["wsid1", "wsid2"]
+}
+```
+
+`missing_manifests` lists mods we couldn't analyze because `files_manifest_built=false`. The frontend can show a banner ("X mods haven't been re-fetched since this feature shipped — file conflicts unavailable for them"), and re-clicking sort eventually triggers re-parse on workshop updates.
+
+Reuse path: `_build_result_for_job` already loads ModInfos via `_row_to_modinfo` — the conflicts endpoint follows the same load pattern, then calls `scan_file_conflicts(conn, mods)` instead of `sort_mods`.
+
+### 2.5 Frontend (out of scope for this plan)
+
+A follow-up plan can wire a "File conflicts" warnings section. For now `/api/conflicts` is consumable from curl and lays the groundwork.
+
+---
+
+## 3. Integration B — Content-type detection feeding category derivation
+
+### 3.1 Schema additions
+
+Already covered by §2.1's `mod_parsed` ALTER TABLE (`mod_types` + `files_manifest_built`). One migration file (`init/09_mod_files.sql`) ships both A and B because they share the worker walk.
+
+### 3.2 Worker changes
+
+Folded into §2.2's single-pass walk. No additional file I/O.
+
+### 3.3 New module: `api/categorize.py`
+
+```python
+def types_to_category(mod_types: list[str], name: str) -> str | None:
+    """First mod_type that maps to a sortof CATEGORY_ORDER bucket wins.
+    Returns None if mod_types is empty / Unknown / Dependency-only and we
+    should fall through to the existing derive_category cascade."""
+```
+
+### 3.4 Tag→category mapping (explicit)
+
+| pzmm `mod_type` | sortof `CATEGORY_ORDER` | notes |
+|---|---|---|
+| `Maps`         | `map`        | already covered by `mod.maps non-empty`; types-derived is a fallback |
+| `Vehicles`     | `vehicle`    | name regex `"spawn zone"` already routes to `vehicle_spawn` upstream |
+| `Weapons`      | `weapon`     | wins over `Items` (pzmm prefers list ordering) |
+| `Items`        | *skip*       | too generic — almost every mod has Items; would mis-trigger |
+| `Clothing`     | `wearable`   | armor name-hint check still runs after, can override to `armor` |
+| `Traits`       | `code`       | no dedicated `trait` bucket; `code` is the gameplay-axis fallback |
+| `Professions`  | `profession` | |
+| `Recipes`      | `crafting`   | |
+| `Tiles`        | `tile`       | |
+| `Textures`     | `texture`    | |
+| `Sounds`       | `sound`      | already handled by `Audio` ws_tag; types-derived is a fallback |
+| `Animations`   | *skip*       | no bucket; falls through |
+| `UI`           | `ui`         | |
+| `Translations` | `translation` | |
+| `Lua`          | *skip*       | too generic; falls through |
+| `Patch`        | `patch`      | already detected by `_PATCH_NAME_RE`; types-derived is a fallback |
+| `Dependency`   | `tweaks`     | maps to existing `lib` pill |
+| `Framework`    | `tweaks`     | same |
+| `Unknown`      | *skip*       | falls through |
+
+"*skip*" means: don't return a category; let `derive_category` continue its cascade.
+
+### 3.5 `derive_category` integration
+
+Insert a single new check in `api/mlos_sort.py:derive_category` after the explicit-category early return at line 412, **before** the patch/lib name regex at lines 416–419:
+
+```python
+if mod.mod_types:
+    cat = types_to_category(mod.mod_types, name)
+    if cat:
+        return cat
+```
+
+`mod.mod_types` is added to the `ModInfo` dataclass (`mlos_sort.py:113`). `_row_to_modinfo` (`api/app.py:176`) is updated to read the new column. **Both `mlos_sort.py` copies must change in lockstep.**
+
+**Position rationale:** `mod_types` comes from media-content fingerprinting, more reliable than name regex but less reliable than an explicit `category=` field in `mod.info`. So it sits between (1) explicit category and (2) name regex. The patch/lib regexes that come after still win for true patches/libraries (they'd usually return `Patch`/`Dependency` from detect_mod_types anyway, but we want the regex to win for cases where a "patch mod" hasn't shipped enough media to fingerprint).
+
+Empty `mod_types` (e.g. older rows where `files_manifest_built=false`) means the new check returns `None` and the existing cascade runs unchanged. **Graceful degradation is built in.**
+
+---
+
+## 4. Blockers / risks
+
+### 4.1 Schema migration cost
+- Current cache: **3,123 `workshop_meta` rows, 3,298 `mod_parsed` rows**.
+- New `mod_files` rows estimate: median mod ships ~50 conflict-eligible files (light mods 5–10, heavy framework/map mods 200–500). At 50 avg × 3,298 mods = **~165 k rows**. With sha1 (40 chars) + rel_path (avg 80 chars) + overhead ≈ 200 bytes/row, that's ~33 MB before indexes. Postgres handles this trivially.
+- `ALTER TABLE mod_parsed ADD COLUMN mod_types TEXT[]` and `files_manifest_built BOOLEAN` are additive and metadata-only on Postgres 16 (no rewrite). Instant.
+
+### 4.2 Backfill feasibility
+- The `/tmp/sortof_steam_throttle` flock + `/tmp/sortof_steam_cooldown` 1h kill-switch (worker.py — `fetch_required_wsids`) protect us from Steam metadata 429s. **DD itself does not hit the metadata API**; it hits Steam content servers, which are not part of the rate-limited path. So mass re-DD does not trip the cooldown.
+- Mass re-DD still costs real time: typical DD pull is 20–60 s wall-clock. 3,123 wsids × 30 s avg ÷ 4 drains = **~6.5 hours wall-clock for a full backfill**. Doable but disruptive.
+- **Recommendation: do not run a bulk backfill.** Let the cache populate organically — every workshop update bumps `time_updated`, which triggers a re-parse and now also a manifest build. The `missing_manifests` field in `/api/conflicts` and the empty-`mod_types` graceful-degrade path together mean the feature works on day 1 (empty results for old rows) and improves as authors push updates.
+- Per-mod manual trigger pattern still works (operator-only):
+  ```sql
+  DELETE FROM mod_parsed WHERE workshop_id='<wsid>';
+  INSERT INTO download_jobs (workshop_id, status) VALUES ('<wsid>','queued');
+  ```
+
+### 4.3 Inline detection at sort time
+- Rejected. `detect_mod_types` reads up to ~11 MB per mod from disk (lua/script blobs). With the tempdir destroyed (the actual case), we'd need to re-DD inline — minutes per sort.
+- **All detection runs at parse time** in `process_one`. `derive_category` and `/api/conflicts` are pure DB reads.
+
+---
+
+## 5. Files touched (summary)
+
+**New:**
+- `init/09_mod_files.sql` — `mod_files` table, `mod_parsed.mod_types`, `mod_parsed.files_manifest_built`
+- `api/diagnostics.py` — port of `scan_file_conflicts`, `FileConflict` dataclass
+- `api/categorize.py` — `types_to_category` helper
+
+**Modified:**
+- `worker/worker.py` — extend `process_one`'s `with` block: single-pass walk, manifest + detect_mod_types, upsert rows
+- `worker/worker.py` (top-level) — port `detect_mod_types` from pzmm `mods.py:57–145` (sortof-side copy; do not import from pzmm at runtime)
+- `api/mlos_sort.py` — add `mod_types: List[str]` to `ModInfo` dataclass; add `mod_types` check at top of `derive_category`
+- `worker/mlos_sort.py` — mirror the `ModInfo` and `derive_category` change (worker/api dual-edit rule)
+- `api/app.py` — `_row_to_modinfo` reads new `mod_types` column; `_build_result_for_job` SELECT list adds `mp.mod_types`; register `POST /api/conflicts`
+
+**Out of scope (deferred to follow-up plan):**
+- Frontend conflicts panel — `/api/conflicts` endpoint only, no UI
+- Integration of `pzmm/core/bundle.py` (debug bundle export) — read for context, not ported
+- Backfill orchestration — relying on organic backfill
+
+---
+
+## 6. Rollback
+
+Before applying the migration:
+
+```bash
+# Backup mod_parsed (the only existing table we ALTER)
+sudo docker exec -i sortof_db pg_dump -U sortof -d sortof -t mod_parsed \
+  > /opt/sortof/backups/mod_parsed-pre-09.sql.$(date +%Y%m%d-%H%M)
+ls -la /opt/sortof/backups/ | tail -3
+```
+
+Down SQL (paste into psql to revert the schema half of this plan):
+
+```sql
+DROP TABLE IF EXISTS mod_files;
+ALTER TABLE mod_parsed
+  DROP COLUMN IF EXISTS mod_types,
+  DROP COLUMN IF EXISTS files_manifest_built;
+```
+
+To revert code, `git checkout main` and restart services:
+```bash
+sudo systemctl restart sortof-api sortof-drain@1 sortof-drain@2 sortof-drain@3 sortof-drain@4
+```
+
+The migration is additive only (new table + new columns with safe defaults), so the rollback is a clean drop. No data is destroyed in `mod_parsed`'s existing columns.
+
+---
+
+## 7. Verification
+
+1. **Migration applies cleanly:**
+   ```bash
+   sudo docker exec -i sortof_db psql -U sortof -d sortof < /opt/sortof/init/09_mod_files.sql
+   sudo docker exec -i sortof_db psql -U sortof -d sortof -c "\d mod_files"
+   sudo docker exec -i sortof_db psql -U sortof -d sortof -c "\d mod_parsed" | grep -E "mod_types|files_manifest_built"
+   ```
+
+2. **Compile checks** (after every Python edit):
+   ```bash
+   /opt/sortof/api/.venv/bin/python -m py_compile /opt/sortof/api/app.py /opt/sortof/api/mlos_sort.py /opt/sortof/api/diagnostics.py /opt/sortof/api/categorize.py
+   /opt/sortof/worker/.venv/bin/python -m py_compile /opt/sortof/worker/worker.py /opt/sortof/worker/mlos_sort.py
+   cd /opt/sortof/api    && .venv/bin/python -c "import app"   && echo OK
+   cd /opt/sortof/worker && .venv/bin/python -c "import drain" && echo OK
+   ```
+
+3. **Dual-edit consistency check** (worker/api `mlos_sort.py` lockstep rule):
+   ```bash
+   diff /opt/sortof/api/mlos_sort.py /opt/sortof/worker/mlos_sort.py | grep -E "^[<>]" | head -20
+   ```
+   Logic must match; only comments / docstrings may differ. If any logic line shows up in the diff, fix the lockstep before continuing.
+
+4. **Restart services:**
+   ```bash
+   sudo systemctl restart sortof-api sortof-drain@1 sortof-drain@2 sortof-drain@3 sortof-drain@4
+   sudo systemctl is-active sortof-api sortof-drain@{1..4}
+   ```
+
+5. **Force a fresh parse on a known multi-file mod and verify manifest:**
+   ```bash
+   sudo docker exec -i sortof_db psql -U sortof -d sortof -c \
+     "DELETE FROM mod_parsed WHERE workshop_id='2169435993';
+      INSERT INTO download_jobs (workshop_id, status) VALUES ('2169435993','queued');"
+   sleep 60
+   sudo docker exec -i sortof_db psql -U sortof -d sortof -c \
+     "SELECT mod_id, mod_types, files_manifest_built FROM mod_parsed WHERE workshop_id='2169435993';
+      SELECT count(*) AS file_count FROM mod_files WHERE workshop_id='2169435993';"
+   ```
+   Expected: `files_manifest_built=t`, `mod_types` populated, `file_count > 0`.
+
+6. **Conflict endpoint smoke:**
+   ```bash
+   curl -sS -X POST http://100.114.205.53:8801/api/conflicts \
+     -H 'Content-Type: application/json' \
+     -d '{"input":"2169435993;2392709985;2487022075"}' | jq .
+   ```
+   Expected: `{"conflicts": [], "missing_manifests": [<wsids without manifests yet>]}`.
+
+7. **Collection-input rejection (Q4):**
+   ```bash
+   curl -sS -i -X POST http://100.114.205.53:8801/api/conflicts \
+     -H 'Content-Type: application/json' \
+     -d '{"input":"https://steamcommunity.com/sharedfiles/filedetails/?id=999999999"}' | head -5
+   ```
+   Expected: HTTP 400 with the documented `detail` message (when the URL is detected as a collection ref).
+
+8. **Category-from-types smoke:**
+   - Find a mod whose Steam tags don't reflect content (e.g. weapon mod tagged only `Realistic`); `/api/sort` currently classifies it as `code` / `other` / `undefined`.
+   - Re-queue it through the new pipeline (delete+insert).
+   - Re-run `/api/sort`; confirm category is now `weapon`.
+
+9. **Graceful-degradation check:** confirm a mod with `files_manifest_built=false` still sorts correctly through the existing cascade (no exceptions, category falls back to current behavior).

init/09_mod_files.sql

@@ -0,0 +1,34 @@
+-- pzmm-conflict-typing migration (Plan: docs/plans/2026-05-04-pzmm-conflict-and-typing.md)
+--
+-- Adds:
+--   - mod_files: per-mod manifest of conflict-eligible asset files (lua/txt/xml/json/ini)
+--                with sha1 fingerprint. Used by /api/conflicts to flag rel_paths claimed
+--                by ≥2 mods with non-equal content.
+--   - mod_parsed.mod_types: ordered tag list from detect_mod_types content fingerprinting
+--                (Maps, Vehicles, Weapons, Traits, …). Consumed by derive_category.
+--   - mod_parsed.files_manifest_built: graceful-degradation flag. False until the
+--                worker has run the single-pass media/ walk that produces both the
+--                manifest rows and mod_types. Old cached rows stay false until they
+--                organically re-parse.
+--
+-- Backfill strategy: organic only — time_updated bumps trigger re-parse, which
+-- now also runs the manifest walk. /api/conflicts surfaces missing_manifests so
+-- the frontend can communicate gaps without lying.
+
+CREATE TABLE IF NOT EXISTS mod_files (
+    workshop_id  TEXT NOT NULL,
+    mod_id       TEXT NOT NULL,
+    rel_path     TEXT NOT NULL,
+    sha1         TEXT NOT NULL,
+    size_bytes   INTEGER NOT NULL DEFAULT 0,
+    PRIMARY KEY (workshop_id, mod_id, rel_path),
+    FOREIGN KEY (workshop_id, mod_id)
+        REFERENCES mod_parsed (workshop_id, mod_id) ON DELETE CASCADE
+);
+
+CREATE INDEX IF NOT EXISTS mod_files_rel_path_idx ON mod_files (rel_path);
+CREATE INDEX IF NOT EXISTS mod_files_mod_idx ON mod_files (workshop_id, mod_id);
+
+ALTER TABLE mod_parsed
+    ADD COLUMN IF NOT EXISTS mod_types            TEXT[]  NOT NULL DEFAULT '{}',
+    ADD COLUMN IF NOT EXISTS files_manifest_built BOOLEAN NOT NULL DEFAULT FALSE;

worker/mlos_sort.py

@@ -130,6 +130,11 @@ class ModInfo:
     # signal for build / multiplayer / category detection. Distinct from
     # `tags` which is mod.info-side (freeform).
     workshop_tags: List[str] = field(default_factory=list)
+    # pzmm-style content fingerprint (Maps, Vehicles, Weapons, Traits, …)
+    # populated by worker.build_manifest_and_types at parse time. Empty when
+    # files_manifest_built=false (older cached rows); derive_category falls
+    # through to the existing cascade in that case.
+    mod_types: List[str] = field(default_factory=list)
     warnings: Dict[str, List[str]] = field(default_factory=dict)
 
 
@@ -347,8 +352,15 @@ def load_mods_from_dir(root: Path) -> List[ModInfo]:
 _PATCH_NAME_RE = re.compile(r"\b(patch|compat|compatibility)\b", re.IGNORECASE)
 
 
-# Substring-based category hints (kept in sync with api/mlos_sort.py)
-_LIB_NAME_HINTS  = ["library", "libraries", "framework"]
+# Substring lists used for derive_category name heuristics. Plain substring
+# matching (vs. \b regex) survives PZ's mishmash of camelCase + underscore
+# + version-suffix mod names (TrueActions_1.09, TrueMusic, TMMumble, …)
+# that strict word boundaries fail on. False positives are accepted in
+# exchange — names containing "music" without being music-related are rare
+# in PZ.
+_LIB_NAME_HINTS = [
+    "library", "libraries", "framework",
+]
 _LIB_NAME_RE = re.compile(
     r'(?<![A-Za-z])(?:lib|api|core)(?![A-Za-z])'
     r'|(?<=[a-z])(?:Lib|API|Core)(?![A-Za-z])',
@@ -381,6 +393,43 @@ def _name_has(name: str, hints: List[str]) -> bool:
     return any(h in n for h in hints)
 
 
+# pzmm content-type → sortof CATEGORY_ORDER mapping. "skip" entries fall
+# through to the existing derive_category cascade. Items/Animations/Lua/Unknown
+# are too generic; Maps/Sounds/Patch/Vehicles/Clothing duplicate signals already
+# captured by the cascade but stay here as fallbacks for poorly-tagged mods.
+_TYPE_TO_CAT: Dict[str, str] = {
+    "Maps":         "map",
+    "Vehicles":     "vehicle",
+    "Weapons":      "weapon",
+    "Clothing":     "wearable",
+    "Traits":       "code",
+    "Professions":  "profession",
+    "Recipes":      "crafting",
+    "Tiles":        "tile",
+    "Textures":     "texture",
+    "Sounds":       "sound",
+    "UI":           "ui",
+    "Translations": "translation",
+    "Patch":        "patch",
+    "Dependency":   "tweaks",
+    "Framework":    "tweaks",
+}
+
+
+def _types_to_category(mod_types: List[str], name: str) -> Optional[str]:
+    """First mod_type that maps to a sortof CATEGORY_ORDER bucket wins.
+    Returns None if mod_types contains only skip-types (Items/Animations/Lua/
+    Unknown), so the caller can fall through to the existing cascade."""
+    for t in mod_types:
+        cat = _TYPE_TO_CAT.get(t)
+        if cat:
+            # vehicle_spawn refinement matches the downstream ws_tag check.
+            if cat == "vehicle" and name and "spawn zone" in name.lower():
+                return "vehicle_spawn"
+            return cat
+    return None
+
+
 def derive_category(mod: ModInfo) -> str:
     """Best-effort category from mod.info + workshop_meta.tags + name.
     Mirrors api/mlos_sort.py; keep both copies in sync.
@@ -389,6 +438,15 @@ def derive_category(mod: ModInfo) -> str:
         return mod.category
 
     name = mod.name or ""
+
+    # pzmm-style content fingerprint takes precedence over name regex when
+    # available. Empty mod_types means files_manifest_built=false (older
+    # cached row); fall through to existing cascade.
+    if mod.mod_types:
+        cat = _types_to_category(mod.mod_types, name)
+        if cat:
+            return cat
+
     if name and _PATCH_NAME_RE.search(name):
         return "patch"
     if _name_has(name, _LIB_NAME_HINTS) or (name and _LIB_NAME_RE.search(name)):

worker/worker.py

@@ -19,6 +19,7 @@ from __future__ import annotations
 
 import argparse
 import asyncio
+import hashlib
 import json
 import os
 import re
@@ -27,7 +28,7 @@ import subprocess
 import sys
 import tempfile
 from pathlib import Path
-from typing import Dict, List, Optional
+from typing import Dict, List, Optional, Tuple
 
 import asyncpg
 import httpx
@@ -208,14 +209,21 @@ def run_depot_downloader(
     workshop_id: str,
     output_dir: Path,
     dd_path: Path,
-    filelist_regex: str = r"regex:.*\.info$",
+    # mod.info / map.info for the existing parse path, plus _CONFLICT_EXTS
+    # (lua/txt/xml/json/ini) for build_manifest_and_types — both the conflict
+    # manifest and pzmm-style content sniffing live on these. Binary assets
+    # (.png/.dds/.bank/.ogg/.wav/.X) are intentionally excluded; their type
+    # signals degrade gracefully via workshop_meta.tags fallback in
+    # mlos_sort.derive_category.
+    filelist_regex: str = r"regex:.*\.(info|lua|txt|xml|json|ini)$",
     timeout: int = 300,
     max_attempts: int = 3,
     backoff_s: float = 2.0,
 ) -> bool:
     """
-    Fetch workshop item using DepotDownloader, filtered to .info files only.
-    Writes <output_dir>/mods/<mod_id>/mod.info (and possibly map.info paths).
+    Fetch workshop item using DepotDownloader, filtered to mod.info plus
+    conflict-eligible asset files (lua/txt/xml/json/ini). Writes
+    <output_dir>/mods/<mod_id>/mod.info (and the asset tree under media/).
     Returns True on success.
 
     Retries up to max_attempts times on rc!=0 or timeout - Steam Workshop's
@@ -308,6 +316,178 @@ def discover_map_folders(mip_parent: Path) -> List[str]:
     return out
 
 
+# -----------------------------------------------------------------------------
+# Single-pass manifest + content-type detection (Plan: 2026-05-04 pzmm port)
+#
+# Both Integration A (file-conflict manifest) and Integration B (mod_types
+# fingerprinting) read the same files from the temp DD extraction. We walk
+# the mod_id root once, hashing conflict-eligible files into a manifest map
+# AND collecting pzmm-style content signals into a tag set in the same loop.
+# No two-pass implementations.
+# -----------------------------------------------------------------------------
+
+_CONFLICT_EXTS = {".lua", ".txt", ".xml", ".json", ".ini"}
+
+# map subtrees are namespaced per-mod by directory; conflict surface is ~zero,
+# and worldmap.xml can be tens of MB. Skipping anything under maps/<MapName>/
+# from manifest insertion AND mod_types content sniffing. map.info itself
+# still drives `mod.maps` (via discover_map_folders) which feeds the existing
+# `mod.maps non-empty → map` rule in derive_category.
+_RE_MAP_SUBTREE = re.compile(r"^maps/[^/]+/")
+
+# Ordered by pzmm preference: first match wins when types_to_category maps
+# this list down to a single sortof CATEGORY_ORDER bucket.
+_TYPE_PREFERRED = [
+    "Maps", "Vehicles", "Weapons", "Items", "Clothing", "Traits",
+    "Professions", "Recipes", "Tiles", "Textures", "Sounds",
+    "Animations", "UI", "Translations", "Lua", "Patch", "Dependency",
+    "Framework",
+]
+
+
+def _sha1(path: Path) -> str:
+    h = hashlib.sha1()
+    with path.open("rb") as f:
+        for chunk in iter(lambda: f.read(1024 * 128), b""):
+            h.update(chunk)
+    return h.hexdigest()
+
+
+def _read_small_text(path: Path, limit: int = 256_000) -> str:
+    try:
+        with path.open("rb") as f:
+            data = f.read(limit)
+        return data.decode("utf-8", errors="ignore").lower()
+    except Exception:
+        return ""
+
+
+def build_manifest_and_types(
+    mip_parent: Path, mod_id: str, raw: str
+) -> Tuple[List[Tuple[str, str, int]], List[str]]:
+    """Single-pass walk under the mod_id root.
+
+    Returns:
+        manifest_rows: list of (rel_path, sha1, size_bytes) for conflict-eligible
+                       files (lua/txt/xml/json/ini). rel_path is lowercased,
+                       posix-style, prefixed with "media/" so it's comparable
+                       across mods regardless of B41/B42/split branch layout.
+                       Last-wins on duplicate rel_paths from multi-branch mods.
+        mod_types:    pzmm-ordered content tag list (Maps, Vehicles, …).
+                      Falls back to ["Dependency"] or ["Unknown"] when no
+                      media/ subtree exists.
+    """
+    if mip_parent.parent.name == "mods":
+        modid_root = mip_parent
+    else:
+        modid_root = mip_parent.parent
+
+    tags: set = set()
+
+    name_blob = (raw or "").lower()
+    if any(w in name_blob for w in ("compat", "compatibility", "patch", "fix")):
+        tags.add("Patch")
+    if any(w in name_blob for w in ("api", "core", "dependency", "framework", "library", "required")):
+        tags.add("Dependency")
+
+    manifest_map: Dict[str, Tuple[str, int]] = {}
+    script_text_parts: List[str] = []
+    lua_text_parts: List[str] = []
+    has_media = False
+
+    for media_dir in modid_root.rglob("media"):
+        if not media_dir.is_dir():
+            continue
+        has_media = True
+        for path in media_dir.rglob("*"):
+            if not path.is_file():
+                continue
+            try:
+                rel_below = path.relative_to(media_dir).as_posix().lower()
+            except ValueError:
+                continue
+            suffix = path.suffix.lower()
+            in_map_subtree = bool(_RE_MAP_SUBTREE.match(rel_below))
+
+            # Manifest: skip per-map subtree (see _RE_MAP_SUBTREE comment).
+            if suffix in _CONFLICT_EXTS and not in_map_subtree:
+                rel = "media/" + rel_below
+                try:
+                    size = path.stat().st_size
+                    sha = _sha1(path)
+                    manifest_map[rel] = (sha, size)
+                except OSError:
+                    pass
+
+            # Path-based mod_type signals — always fire, even for files inside
+            # the per-map subtree. The "Maps" tag is the canonical signal that
+            # this mod ships a map and shouldn't be lost when we exclude the
+            # heavy worldmap.xml / objects.lua content from the manifest.
+            if rel_below.startswith("maps/"):
+                tags.add("Maps")
+            if rel_below.startswith("texturepacks/") or "tiledefinitions" in rel_below:
+                tags.add("Tiles")
+            if (rel_below.startswith(("textures/", "models_x/", "models/"))
+                    or suffix in {".png", ".dds"}):
+                if path.name.lower() != "poster.png":
+                    tags.add("Textures")
+            if rel_below.startswith(("scripts/vehicles/", "scripts/vehicle")):
+                tags.add("Vehicles")
+            if rel_below.startswith(("clothing/", "scripts/clothing/")):
+                tags.add("Clothing")
+            if (rel_below.startswith(("sound/", "sounds/", "fmod/"))
+                    or suffix in {".bank", ".ogg", ".wav"}):
+                tags.add("Sounds")
+            if rel_below.startswith(("ui/", "lua/client/ui/")):
+                tags.add("UI")
+            if rel_below.startswith(("anims/", "animsets/", "actiongroups/")):
+                tags.add("Animations")
+            if rel_below.startswith("lua/shared/translate/") or "/translate/" in rel_below:
+                tags.add("Translations")
+
+            # Content-blob accumulation: skip per-map subtree. Map-internal lua
+            # / scripts (rare but possible) wouldn't normally collide with
+            # other mods anyway, and reading them just costs memory for no
+            # detection upside.
+            if not in_map_subtree:
+                if rel_below.startswith("lua/"):
+                    tags.add("Lua")
+                    if suffix == ".lua" and len(lua_text_parts) < 60:
+                        lua_text_parts.append(_read_small_text(path, 64_000))
+                if (rel_below.startswith("scripts/") and suffix in {".txt", ".xml"}
+                        and len(script_text_parts) < 80):
+                    script_text_parts.append(_read_small_text(path, 96_000))
+
+    if not has_media:
+        return [], (["Dependency"] if "Dependency" in tags else ["Unknown"])
+
+    script_blob = "\n".join(script_text_parts)
+    lua_blob = "\n".join(lua_text_parts)
+
+    if " type = weapon" in script_blob or "displaycategory = weapon" in script_blob:
+        tags.add("Weapons")
+    if " vehicle " in script_blob or "module vehicles" in script_blob:
+        tags.add("Vehicles")
+    if "item " in script_blob:
+        tags.add("Items")
+    if "recipe " in script_blob or " evolvedrecipe " in script_blob:
+        tags.add("Recipes")
+    if "bodylocation" in script_blob or "clothingitem" in script_blob:
+        tags.add("Clothing")
+    if "traitfactory.addtrait" in lua_blob:
+        tags.add("Traits")
+    if "professionfactory.addprofession" in lua_blob:
+        tags.add("Professions")
+
+    has_require = bool(re.search(r"^\s*require\s*=", raw or "", re.IGNORECASE | re.MULTILINE))
+    if not tags and has_require and not script_blob:
+        tags.add("Framework")
+
+    ordered = [t for t in _TYPE_PREFERRED if t in tags] or ["Unknown"]
+    manifest_rows = [(rel, sha, size) for rel, (sha, size) in sorted(manifest_map.items())]
+    return manifest_rows, ordered
+
+
 # -----------------------------------------------------------------------------
 # DB upserts
 # -----------------------------------------------------------------------------
@@ -356,8 +536,9 @@ INSERT INTO mod_parsed (
     requirements, load_after, load_before, incompatible_mods,
     load_first, load_last, tags, maps,
     raw_mod_info, version_min, is_addon,
+    mod_types, files_manifest_built,
     parsed_at_time_updated, parsed_at
-) VALUES ($1,$2,$3,$4,$5,$6,$7,$8,$9,$10,$11,$12,$13,$14,$15,$16, now())
+) VALUES ($1,$2,$3,$4,$5,$6,$7,$8,$9,$10,$11,$12,$13,$14,$15,$16,$17,$18, now())
 ON CONFLICT (workshop_id, mod_id) DO UPDATE SET
     name                   = EXCLUDED.name,
     category               = EXCLUDED.category,
@@ -372,10 +553,24 @@ ON CONFLICT (workshop_id, mod_id) DO UPDATE SET
     raw_mod_info           = EXCLUDED.raw_mod_info,
     version_min            = EXCLUDED.version_min,
     is_addon               = EXCLUDED.is_addon,
+    mod_types              = EXCLUDED.mod_types,
+    files_manifest_built   = EXCLUDED.files_manifest_built,
     parsed_at_time_updated = EXCLUDED.parsed_at_time_updated,
     parsed_at              = now();
 """
 
+DELETE_MOD_FILES = """
+DELETE FROM mod_files WHERE workshop_id = $1 AND mod_id = $2;
+"""
+
+INSERT_MOD_FILE = """
+INSERT INTO mod_files (workshop_id, mod_id, rel_path, sha1, size_bytes)
+VALUES ($1, $2, $3, $4, $5)
+ON CONFLICT (workshop_id, mod_id, rel_path) DO UPDATE SET
+    sha1       = EXCLUDED.sha1,
+    size_bytes = EXCLUDED.size_bytes;
+"""
+
 # Description-text heuristic for "this mod is an optional add-on to the
 # primary mod published by the same wsid". Matches:
 #   "Optional add-on: removes ..."   (TMMumble)
@@ -487,6 +682,10 @@ async def process_one(
             if mod is None:
                 continue
             maps = discover_map_folders(mip.parent)
+            # Single-pass walk under the mod_id root: produces both the
+            # conflict manifest and the pzmm-style mod_types list. See
+            # build_manifest_and_types.
+            manifest_rows, mod_types = build_manifest_and_types(mip.parent, mod.id, raw)
             # Evict any other wsid's claim on this mod_id before we install
             # ours. Cache invariant: at most one wsid per mod_id, with the
             # most-recent pull winning.
@@ -508,8 +707,19 @@ async def process_one(
                 raw,
                 extract_version_min(raw),
                 detect_is_addon(raw),
+                mod_types,
+                True,           # files_manifest_built
                 time_updated,
             )
+            # Replace any stale manifest rows for this (workshop_id, mod_id)
+            # so a re-parse can't leave behind orphans from a prior layout.
+            await conn.execute(DELETE_MOD_FILES, workshop_id, mod.id)
+            if manifest_rows:
+                await conn.executemany(
+                    INSERT_MOD_FILE,
+                    [(workshop_id, mod.id, rel, sha, size)
+                     for rel, sha, size in manifest_rows],
+                )
             seen_mod_ids.append(mod.id)
 
         # Drop rows for mods that no longer exist in this workshop item