resumed testing and handover improvement

modules/features/graphicalEditor/conditionOperators.py

@@ -257,7 +257,7 @@ def _path_suggests_file(path: List[Any], producer_type: str) -> bool:
     return False
 
 
-def resolve_value_kind(graph: Dict[str, Any], ref: Dict[str, Any]) -> str:
+def resolve_value_kind(graph: Dict[str, Any], ref: Dict[str, Any], *, _skip_upstream: bool = False) -> str:
     """Resolve condition valueKind for a DataRef against the workflow graph."""
     if not isinstance(ref, dict):
         return "unknown"
@@ -281,31 +281,32 @@ def resolve_value_kind(graph: Dict[str, Any], ref: Dict[str, Any]) -> str:
             return "string"
         return "file"
 
-    from modules.features.graphicalEditor.upstreamPathsService import compute_upstream_paths
+    if not _skip_upstream:
+        from modules.features.graphicalEditor.upstreamPathsService import compute_upstream_paths
 
-    target_id = graph.get("targetNodeId") or producer_id
-    matched_type: Optional[str] = None
-    for entry in compute_upstream_paths(graph, target_id):
-        if entry.get("producerNodeId") != producer_id:
-            continue
-        entry_path = entry.get("path") or []
-        if _paths_equal(list(entry_path), list(path)):
-            matched_type = str(entry.get("type") or "Any")
-            break
-
-    if matched_type is None and path:
-        parent_path = list(path[:-1])
+        target_id = graph.get("targetNodeId") or producer_id
+        matched_type: Optional[str] = None
         for entry in compute_upstream_paths(graph, target_id):
             if entry.get("producerNodeId") != producer_id:
                 continue
-            if _paths_equal(list(entry.get("path") or []), parent_path):
+            entry_path = entry.get("path") or []
+            if _paths_equal(list(entry_path), list(path)):
                 matched_type = str(entry.get("type") or "Any")
                 break
 
-    if matched_type:
-        vk = catalog_type_to_value_kind(matched_type)
-        if vk != "unknown":
-            return vk
+        if matched_type is None and path:
+            parent_path = list(path[:-1])
+            for entry in compute_upstream_paths(graph, target_id):
+                if entry.get("producerNodeId") != producer_id:
+                    continue
+                if _paths_equal(list(entry.get("path") or []), parent_path):
+                    matched_type = str(entry.get("type") or "Any")
+                    break
+
+        if matched_type:
+            vk = catalog_type_to_value_kind(matched_type)
+            if vk != "unknown":
+                return vk
 
     if producer_type in ("trigger.form", "input.form") and path and str(path[0]) == "payload":
         return "string"
@@ -414,6 +415,12 @@ def _iter_presentation_parts(envelope: Dict[str, Any]) -> List[Dict[str, Any]]:
         if not isinstance(bucket, dict):
             continue
         data = bucket.get("data")
+        mode = str(bucket.get("outputMode") or "").strip().lower()
+        if mode == "blob" and isinstance(data, str):
+            from modules.workflows.methods.methodContext.actions.extractContent import parse_blob_data_segments
+
+            parts.extend(parse_blob_data_segments(data))
+            continue
         if isinstance(data, list):
             for slot in data:
                 if isinstance(slot, dict):

modules/features/graphicalEditor/nodeDefinitions/flow.py

@@ -123,9 +123,27 @@ CONTEXT_MERGE_ACTION_RESULT_DATA_PICK_OPTIONS = [
     },
 ]
 
+_CONTEXT_BRANCH_DATA_PICK_OPTIONS = [
+    {
+        "path": ["items"],
+        "pickerLabel": t("Gefilterte Elemente"),
+        "detail": t("Empfohlen für Schleifen: je Eintrag ein Durchlauf (z. B. Bild-Slots)."),
+        "recommended": True,
+        "type": "List[Any]",
+    },
+    {
+        "path": ["data"],
+        "pickerLabel": t("Kontext (data)"),
+        "detail": t("Gefilterter Presentation-Umschlag oder unveränderter Eingang auf dem Sonst-Zweig."),
+        "recommended": False,
+        "type": "Dict",
+    },
+]
+
 # Ports, die typische Schritt-Ausgaben durchreichen (nicht nur leerer Transit).
 _FLOW_INPUT_SCHEMAS = [
     "Transit",
+    "ContextBranch",
     "FormPayload",
     "AiResult",
     "TextResult",
@@ -183,8 +201,10 @@ FLOW_NODES = [
         "category": "flow",
         "label": t("Switch"),
         "description": t(
-            "Mehrere Zweige nach einem Wert aus einem vorherigen Schritt (Data Picker). "
-            "Definiere Fälle mit Vergleichsoperator; der Eingang wird an den ersten passenden Zweig durchgereicht."
+            "Mehrere Zweige nach einem Wert aus einem vorherigen Schritt. "
+            "Jeder Fall hat einen eigenen Ausgang mit passend gefiltertem Inhalt in ``items``; "
+            "mehrere Kontext-Filter können gleichzeitig zutreffen (z. B. Text und Bilder). "
+            "Der letzte Ausgang (Sonst) reicht den unveränderten Eingang durch."
         ),
         "parameters": [
             {
@@ -199,13 +219,22 @@ FLOW_NODES = [
                 "type": "array",
                 "required": False,
                 "frontendType": "caseList",
-                "description": t("Fälle: Operator und Vergleichswert"),
+                "frontendOptions": {
+                    "dependsOn": "value",
+                    "operatorCatalog": "condition",
+                },
+                "description": t("Fälle: Operator und Vergleichswert (abhängig vom gewählten Wert)"),
             },
         ],
         "inputs": 1,
         "outputs": 1,
         "inputPorts": {0: {"accepts": list(_FLOW_INPUT_SCHEMAS)}},
-        "outputPorts": {0: {"schema": "Transit"}},
+        "outputPorts": {
+            0: {
+                "schema": "ContextBranch",
+                "dataPickOptions": _CONTEXT_BRANCH_DATA_PICK_OPTIONS,
+            },
+        },
         "executor": "flow",
         "meta": {"icon": "mdi-swap-horizontal", "color": "#FF9800", "usesAi": False},
     },
@@ -265,7 +294,7 @@ FLOW_NODES = [
         "outputLabels": [t("Schleife"), t("Fertig")],
         "inputPorts": {
             0: {"accepts": [
-                "Transit", "UdmDocument", "EmailList", "DocumentList", "FileList", "TaskList",
+                "Transit", "ContextBranch", "UdmDocument", "EmailList", "DocumentList", "FileList", "TaskList",
                 "ActionResult", "AiResult", "QueryResult", "FormPayload", "LoopItem",
             ]},
         },

modules/features/graphicalEditor/portTypes.py

@@ -298,6 +298,21 @@ PORT_TYPE_CATALOG: Dict[str, PortSchema] = {
         PortField(name="merged", type="Dict",
                   description="Zusammengeführte Daten"),
     ]),
+    "ContextBranch": PortSchema(name="ContextBranch", fields=[
+        PortField(name="items", type="List[Any]",
+                  description="Schleifen-fertige Elemente aus dem (gefilterten) Kontext",
+                  recommended=True,
+                  picker_label=t("Gefilterte Elemente")),
+        PortField(name="data", type="Dict", required=False,
+                  description="Gefilterter Presentation-Umschlag oder Eingabe-Spiegel",
+                  picker_label=t("Kontext (data)")),
+        PortField(name="filterApplied", type="bool", required=False,
+                  description="True wenn ein Kontext-Inhaltsfilter angewendet wurde"),
+        PortField(name="contentType", type="str", required=False,
+                  description="Angewendeter Inhaltstyp-Filter (z. B. image)"),
+        PortField(name="match", type="int", required=False,
+                  description="Aktiver Ausgangs-Index (Fall oder Sonst)"),
+    ]),
     "ActionDocument": PortSchema(name="ActionDocument", fields=[
         PortField(name="documentName", type="str",
                   description="Dokumentname",

modules/features/graphicalEditor/switchOutput.py

@@ -0,0 +1,308 @@
+# Copyright (c) 2025 Patrick Motsch
+"""Build flow.switch branch payloads: filtered context + loop-ready items."""
+
+from __future__ import annotations
+
+import copy
+import re
+from typing import Any, Dict, List, Optional
+
+from modules.features.graphicalEditor.portTypes import unwrapTransit
+
+_CONTEXT_FILTER_OPERATORS = frozenset({"contains_content"})
+_BLOB_IMAGE_CHUNK_RE = re.compile(r"^\[image(?:\:([^\]]+))?\]$")
+
+
+def _artifacts_by_part_id_from_presentation(inp: Any) -> Dict[str, str]:
+    plain = _unwrap_input(inp)
+    meta = plain.get("_meta") if isinstance(plain, dict) else None
+    if not isinstance(meta, dict):
+        return {}
+    out: Dict[str, str] = {}
+    for art in meta.get("persistedImageArtifacts") or []:
+        if not isinstance(art, dict):
+            continue
+        sp = str(art.get("sourcePartId") or "").strip()
+        fid = str(art.get("fileId") or "").strip()
+        if sp and fid:
+            out[sp] = fid
+    return out
+
+
+def _enrich_image_slot(slot: Dict[str, Any], artifacts_by_part: Dict[str, str]) -> None:
+    if (slot.get("typeGroup") or "").strip().lower() != "image":
+        return
+    existing = str(slot.get("embeddedImageFileId") or "").strip()
+    if existing and existing in artifacts_by_part.values():
+        return
+    candidates: List[str] = []
+    sid = str(slot.get("id") or "").strip()
+    if sid:
+        candidates.append(sid)
+    data = slot.get("data")
+    if isinstance(data, str):
+        m = _BLOB_IMAGE_CHUNK_RE.fullmatch(data.strip())
+        if m:
+            tok = (m.group(1) or "").strip()
+            if tok:
+                candidates.append(tok)
+    for cand in candidates:
+        fid = artifacts_by_part.get(cand)
+        if fid:
+            slot["embeddedImageFileId"] = fid
+            return
+
+
+def _slot_matches_content_type(slot: Dict[str, Any], content_type: str) -> bool:
+    target = (content_type or "").strip().lower()
+    if not target:
+        return False
+    tg = (slot.get("typeGroup") or slot.get("contentType") or "").strip().lower()
+    if target == "media":
+        return tg in ("image", "media", "video", "audio")
+    if target == "text":
+        return tg in ("text", "table", "structure")
+    return tg == target
+
+
+def _filter_bucket_slots(bucket: Dict[str, Any], content_type: str) -> Dict[str, Any]:
+    """Return a copy of a presentation file bucket with filtered ``data`` slots."""
+    mode = str(bucket.get("outputMode") or "").strip().lower()
+    data = bucket.get("data")
+    if mode == "blob" and isinstance(data, str):
+        from modules.workflows.methods.methodContext.actions.extractContent import (
+            filter_blob_bucket_by_content_type,
+        )
+
+        return filter_blob_bucket_by_content_type(bucket, content_type)
+    out = copy.deepcopy(bucket)
+    if isinstance(data, list):
+        out["data"] = [s for s in data if isinstance(s, dict) and _slot_matches_content_type(s, content_type)]
+    elif isinstance(data, dict) and _slot_matches_content_type(data, content_type):
+        out["data"] = data
+    else:
+        out["data"] = [] if isinstance(data, list) else data
+    return out
+
+
+def _filter_presentation_envelope(envelope: Dict[str, Any], content_type: str) -> Dict[str, Any]:
+    """Filter all slots in a presentation envelope by content type group."""
+    from modules.workflows.methods.methodContext.actions.extractContent import (
+        PRESENTATION_KIND,
+        PRESENTATION_SCHEMA_VERSION,
+    )
+
+    out = copy.deepcopy(envelope)
+    files = out.get("files") or {}
+    if not isinstance(files, dict):
+        return out
+    filtered_files: Dict[str, Any] = {}
+    kept_order: List[str] = []
+    for fk in out.get("fileOrder") or list(files.keys()):
+        bucket = files.get(fk)
+        if not isinstance(bucket, dict):
+            continue
+        fb = _filter_bucket_slots(bucket, content_type)
+        data = fb.get("data")
+        has_data = (
+            (isinstance(data, list) and len(data) > 0)
+            or (isinstance(data, dict))
+            or (isinstance(data, str) and str(data).strip())
+        )
+        if has_data:
+            filtered_files[str(fk)] = fb
+            kept_order.append(str(fk))
+    out["schemaVersion"] = out.get("schemaVersion") or PRESENTATION_SCHEMA_VERSION
+    out["kind"] = out.get("kind") or PRESENTATION_KIND
+    out["fileOrder"] = kept_order
+    out["files"] = filtered_files
+    return out
+
+
+def _slots_from_bucket(bucket: Dict[str, Any]) -> List[Any]:
+    data = bucket.get("data")
+    mode = str(bucket.get("outputMode") or "").strip().lower()
+    if mode == "blob" and isinstance(data, str) and data.strip():
+        from modules.workflows.methods.methodContext.actions.extractContent import parse_blob_data_segments
+
+        return parse_blob_data_segments(data)
+    if isinstance(data, list):
+        return [s for s in data if isinstance(s, dict)]
+    if isinstance(data, dict):
+        return [data]
+    if isinstance(data, str) and data.strip():
+        return [{"typeGroup": "text", "data": data}]
+    items = bucket.get("items")
+    if isinstance(items, list):
+        return [i for i in items if isinstance(i, dict)]
+    return []
+
+
+def _items_from_presentation_envelope(
+    envelope: Dict[str, Any],
+    *,
+    artifacts_by_part: Optional[Dict[str, str]] = None,
+) -> List[Any]:
+    items: List[Any] = []
+    files = envelope.get("files") or {}
+    if not isinstance(files, dict):
+        return items
+    for fk in envelope.get("fileOrder") or list(files.keys()):
+        bucket = files.get(fk)
+        if isinstance(bucket, dict):
+            for slot in _slots_from_bucket(bucket):
+                if artifacts_by_part:
+                    _enrich_image_slot(slot, artifacts_by_part)
+                sid = str(slot.get("id") or slot.get("label") or len(items))
+                items.append({"name": f"{fk}:{sid}", "value": slot})
+    return items
+
+
+def expand_items_from_input(raw: Any) -> List[Any]:
+    """Best-effort loop items from transit/presentation/list/dict input."""
+    if raw is None:
+        return []
+    if isinstance(raw, dict) and isinstance(raw.get("items"), list):
+        return list(raw["items"])
+    plain = unwrapTransit(raw) if isinstance(raw, dict) and raw.get("_transit") else raw
+    if isinstance(plain, dict) and isinstance(plain.get("items"), list):
+        return list(plain["items"])
+    from modules.workflows.methods.methodContext.actions.extractContent import (
+        normalize_presentation_envelopes,
+    )
+
+    envelopes = normalize_presentation_envelopes(plain)
+    if envelopes:
+        out: List[Any] = []
+        for env in envelopes:
+            out.extend(_items_from_presentation_envelope(env))
+        return out
+    if isinstance(plain, list):
+        return list(plain)
+    if isinstance(plain, dict):
+        children = plain.get("children")
+        if isinstance(children, list) and children:
+            return list(children)
+        return [{"name": k, "value": v} for k, v in plain.items()]
+    return [plain]
+
+
+def _unwrap_input(inp: Any) -> Any:
+    if isinstance(inp, dict) and inp.get("_transit"):
+        return unwrapTransit(inp)
+    return inp
+
+
+def build_switch_branch_payload(
+    inp: Any,
+    case: Dict[str, Any],
+    *,
+    value_kind: str = "unknown",
+    match_index: int = 0,
+) -> Dict[str, Any]:
+    """Payload for a matched switch case (ContextBranch inner data)."""
+    operator = str(case.get("operator") or "eq")
+    right = case.get("value")
+    plain_in = _unwrap_input(inp)
+
+    if operator in _CONTEXT_FILTER_OPERATORS and value_kind == "context":
+        content_type = str(right or "")
+        from modules.workflows.methods.methodContext.actions.extractContent import (
+            normalize_presentation_envelopes,
+        )
+
+        source = plain_in
+        if isinstance(source, dict) and "data" in source and not source.get("kind"):
+            nested = source.get("data")
+            if isinstance(nested, dict):
+                source = nested
+        envelopes = normalize_presentation_envelopes(source)
+        if not envelopes and isinstance(plain_in, dict):
+            envelopes = normalize_presentation_envelopes(plain_in)
+        filtered_envs = [_filter_presentation_envelope(env, content_type) for env in envelopes]
+        artifacts_by_part = _artifacts_by_part_id_from_presentation(plain_in)
+        items: List[Any] = []
+        for env in filtered_envs:
+            items.extend(_items_from_presentation_envelope(env, artifacts_by_part=artifacts_by_part))
+        if len(filtered_envs) == 1:
+            data_out: Any = filtered_envs[0]
+        elif filtered_envs:
+            data_out = {"envelopes": filtered_envs}
+        else:
+            data_out = {}
+        return {
+            "data": data_out,
+            "items": items,
+            "filterApplied": True,
+            "contentType": content_type,
+            "match": match_index,
+        }
+
+    data_out = plain_in if isinstance(plain_in, dict) else {"value": plain_in}
+    return {
+        "data": data_out,
+        "items": expand_items_from_input(inp),
+        "filterApplied": False,
+        "match": match_index,
+    }
+
+
+def build_switch_default_payload(inp: Any, *, match_index: int) -> Dict[str, Any]:
+    """Sonst branch: unmodified input passthrough."""
+    plain_in = _unwrap_input(inp)
+    data_out = plain_in if isinstance(plain_in, dict) else {"value": plain_in}
+    return {
+        "data": data_out,
+        "items": expand_items_from_input(inp),
+        "filterApplied": False,
+        "match": match_index,
+    }
+
+
+def build_switch_combined_output(
+    inp: Any,
+    cases: List[Any],
+    *,
+    matched_indices: List[int],
+    value_kind: str = "unknown",
+) -> Dict[str, Any]:
+    """Build per-port branch payloads; primary fields mirror the first active match."""
+    branches: Dict[str, Dict[str, Any]] = {}
+    default_idx = len(cases)
+    for idx in matched_indices:
+        if idx == default_idx:
+            branches[str(idx)] = build_switch_default_payload(inp, match_index=default_idx)
+        elif 0 <= idx < len(cases):
+            c = cases[idx] if isinstance(cases[idx], dict) else {"operator": "eq", "value": cases[idx]}
+            branches[str(idx)] = build_switch_branch_payload(
+                inp, c, value_kind=value_kind, match_index=idx,
+            )
+    primary_idx = matched_indices[0] if matched_indices else default_idx
+    primary = branches.get(str(primary_idx)) or build_switch_default_payload(inp, match_index=default_idx)
+    return {**primary, "branches": branches}
+
+
+def switch_branch_payload(transit: Any, source_output: int) -> Optional[Dict[str, Any]]:
+    """Return the ContextBranch inner dict for a specific switch output port."""
+    if not isinstance(transit, dict):
+        return None
+    data = transit.get("data") if transit.get("_transit") else transit
+    if not isinstance(data, dict):
+        return None
+    branches = data.get("branches")
+    if isinstance(branches, dict):
+        branch = branches.get(str(source_output))
+        if isinstance(branch, dict):
+            return branch
+    if transit.get("_transit"):
+        return data
+    return data
+
+
+def unwrap_transit_for_port(output: Any, source_output: Optional[int] = None) -> Any:
+    """Unwrap transit; when ``source_output`` is set, pick that switch branch payload."""
+    if source_output is not None:
+        branch = switch_branch_payload(output, source_output)
+        if branch is not None:
+            return branch
+    return unwrapTransit(output)

modules/features/graphicalEditor/upstreamPathsService.py

@@ -4,7 +4,7 @@ from __future__ import annotations
 
 from typing import Any, Dict, List, Set
 
-from modules.features.graphicalEditor.conditionOperators import resolve_value_kind
+from modules.features.graphicalEditor.conditionOperators import catalog_type_to_value_kind, resolve_value_kind
 from modules.features.graphicalEditor.nodeDefinitions import STATIC_NODE_TYPES
 from modules.features.graphicalEditor.portTypes import PORT_TYPE_CATALOG, PortSchema, parse_graph_defined_output_schema
 from modules.workflows.automation2.graphUtils import buildConnectionMap, getLoopBodyNodeIds, getLoopDoneNodeIds
@@ -169,12 +169,16 @@ def compute_upstream_paths(graph: Dict[str, Any], target_node_id: str) -> List[D
             )
 
     for entry in paths:
-        ref = {
-            "nodeId": entry.get("producerNodeId"),
-            "path": entry.get("path") or [],
-        }
-        graph_with_target = {**graph, "targetNodeId": target_node_id}
-        entry["valueKind"] = resolve_value_kind(graph_with_target, ref)
+        ct = str(entry.get("type") or "Any")
+        vk = catalog_type_to_value_kind(ct)
+        if vk == "unknown":
+            ref = {
+                "nodeId": entry.get("producerNodeId"),
+                "path": entry.get("path") or [],
+            }
+            graph_with_target = {**graph, "targetNodeId": target_node_id}
+            vk = resolve_value_kind(graph_with_target, ref, _skip_upstream=True)
+        entry["valueKind"] = vk
 
     return paths
 

modules/workflows/automation2/executionEngine.py

@@ -163,10 +163,15 @@ def _is_node_on_active_path(
         meta = out.get("_meta", {}) if out.get("_transit") else out
         branch = meta.get("branch")
         match = meta.get("match")
+        matches = meta.get("matches")
 
         active_output = None
         if branch is not None:
             active_output = branch
+        elif isinstance(matches, list) and matches:
+            if source_output not in matches:
+                return False
+            continue
         elif match is not None:
             if match < 0:
                 return False

modules/workflows/automation2/executors/actionNodeExecutor.py

@@ -475,7 +475,7 @@ def _resolveUpstreamPayload(nodeId: str, context: Dict[str, Any]) -> Any:
     the first ``connectionMap`` entry so ``injectUpstreamPayload`` (e.g.
     ``context.mergeContext`` after ``flow.loop``) still receives data.
     """
-    from modules.features.graphicalEditor.portTypes import unwrapTransit
+    from modules.features.graphicalEditor.switchOutput import unwrap_transit_for_port
 
     nodeOutputs = context.get("nodeOutputs") or {}
     connectionMap = context.get("connectionMap") or {}
@@ -496,25 +496,25 @@ def _resolveUpstreamPayload(nodeId: str, context: Dict[str, Any]) -> Any:
 
     if not entry:
         return None
-    src_node_id, _ = entry
+    src_node_id, src_out = entry
     upstream = nodeOutputs.get(src_node_id)
-    return unwrapTransit(upstream) if isinstance(upstream, dict) else upstream
+    return unwrap_transit_for_port(upstream, src_out)
 
 
 def _resolveBranchInputs(nodeId: str, context: Dict[str, Any]) -> Dict[int, Any]:
     """Return ``Dict[port_index → unwrapped upstream output]`` for every wired input port."""
-    from modules.features.graphicalEditor.portTypes import unwrapTransit
+    from modules.features.graphicalEditor.switchOutput import unwrap_transit_for_port
     src_map = (context.get("inputSources") or {}).get(nodeId) or {}
     nodeOutputs = context.get("nodeOutputs") or {}
     out: Dict[int, Any] = {}
     for port_ix, entry in src_map.items():
         if not entry:
             continue
-        src_node_id, _ = entry
+        src_node_id, src_out = entry
         upstream = nodeOutputs.get(src_node_id)
         if upstream is None:
             continue
-        out[int(port_ix)] = unwrapTransit(upstream) if isinstance(upstream, dict) else upstream
+        out[int(port_ix)] = unwrap_transit_for_port(upstream, src_out)
     return out
 
 
@@ -554,7 +554,12 @@ class ActionNodeExecutor:
         # 1. Resolve parameters (DataRef, SystemVar, Static)
         params = dict(node.get("parameters") or {})
         logger.debug("ActionNodeExecutor node %s raw params keys=%s", nodeId, list(params.keys()))
-        resolvedParams = resolveParameterReferences(params, context.get("nodeOutputs", {}))
+        resolvedParams = resolveParameterReferences(
+            params,
+            context.get("nodeOutputs", {}),
+            consumer_node_id=nodeId,
+            input_sources=context.get("inputSources"),
+        )
         logger.debug("ActionNodeExecutor node %s resolved params keys=%s documentList_present=%s documentList_type=%s", nodeId, list(resolvedParams.keys()), "documentList" in resolvedParams, type(resolvedParams.get("documentList")).__name__)
 
         # 2. Apply defaults from parameter definitions

modules/workflows/automation2/executors/flowExecutor.py

@@ -132,15 +132,24 @@ class FlowExecutor:
         value_kind = "unknown"
         ref_for_kind = left_ref if isinstance(left_ref, dict) else cond.get("ref")
         if isinstance(ref_for_kind, dict) and ref_for_kind.get("nodeId") and node:
-            graph_stub = {
-                "nodes": [{"id": node.get("id"), "type": node.get("type")}],
-                "targetNodeId": node.get("id"),
-            }
+            graph_stub = self._graph_stub_for_ref(node, ref_for_kind, nodeOutputs)
             value_kind = resolve_value_kind(graph_stub, ref_for_kind)
 
         return apply_condition_operator(left, str(operator), right, value_kind)
 
-    def _compare_dates(self, left: Any, right: Any, op) -> bool:
+    def _graph_stub_for_ref(self, node: Dict, ref: Dict, nodeOutputs: Dict) -> Dict[str, Any]:
+        """Minimal graph for ``resolve_value_kind`` (includes value producer when known)."""
+        nodes: List[Dict[str, Any]] = [{"id": node.get("id"), "type": node.get("type")}]
+        producer_id = ref.get("nodeId")
+        if producer_id:
+            ctx = nodeOutputs.get("_context") if isinstance(nodeOutputs.get("_context"), dict) else {}
+            graph_nodes = ctx.get("graphNodesById") if isinstance(ctx.get("graphNodesById"), dict) else {}
+            pnode = graph_nodes.get(producer_id) if isinstance(graph_nodes, dict) else None
+            if isinstance(pnode, dict):
+                nodes.append({"id": producer_id, "type": pnode.get("type", "")})
+            else:
+                nodes.append({"id": producer_id, "type": ""})
+        return {"nodes": nodes, "targetNodeId": node.get("id")}
         """Compare left/right as dates; op(a,b) is the comparison."""
 
         def parse(v):
@@ -197,23 +206,42 @@ class FlowExecutor:
         return bool(resolved)
 
     async def _switch(self, node: Dict, nodeOutputs: Dict, nodeId: str, inputSources: Dict) -> Any:
-        valueExpr = (node.get("parameters") or {}).get("value", "")
+        params = node.get("parameters") or {}
+        valueExpr = params.get("value", "")
         from modules.workflows.automation2.graphUtils import resolveParameterReferences
-        value = resolveParameterReferences(valueExpr, nodeOutputs)
-        cases = (node.get("parameters") or {}).get("cases", [])
-        inp = self._getInputData(nodeId, {nodeId: inputSources}, nodeOutputs)
-        for i, c in enumerate(cases):
-            if self._evalSwitchCase(value, c):
-                return wrapTransit(
-                    unwrapTransit(inp) if inp else inp,
-                    {"match": i, "value": value},
-                )
-        return wrapTransit(
-            unwrapTransit(inp) if inp else inp,
-            {"match": -1, "value": value},
+        from modules.features.graphicalEditor.switchOutput import (
+            build_switch_combined_output,
+            build_switch_default_payload,
         )
 
-    def _evalSwitchCase(self, left: Any, case: Any) -> bool:
+        value = resolveParameterReferences(valueExpr, nodeOutputs)
+        cases = params.get("cases", []) or []
+        value_kind = "unknown"
+        if isinstance(valueExpr, dict) and valueExpr.get("type") == "ref":
+            graph_stub = self._graph_stub_for_ref(node, valueExpr, nodeOutputs)
+            value_kind = resolve_value_kind(graph_stub, valueExpr)
+        inp = self._getInputData(nodeId, {nodeId: inputSources}, nodeOutputs)
+        matched: List[int] = [
+            i for i, c in enumerate(cases)
+            if self._evalSwitchCase(value, c, value_kind=value_kind)
+        ]
+        default_idx = len(cases) if isinstance(cases, list) else 0
+        if not matched:
+            matched = [default_idx]
+        combined = build_switch_combined_output(
+            inp, cases, matched_indices=matched, value_kind=value_kind,
+        )
+        return wrapTransit(
+            combined,
+            {
+                "match": matched[0],
+                "matches": matched,
+                "value": value,
+                "filterApplied": bool(combined.get("filterApplied")),
+            },
+        )
+
+    def _evalSwitchCase(self, left: Any, case: Any, *, value_kind: Optional[str] = None) -> bool:
         """
         Evaluate a switch case. Case can be:
         - dict: {operator, value} - use operator to compare left vs value
@@ -225,14 +253,19 @@ class FlowExecutor:
         else:
             operator = "eq"
             right = case
-        return apply_condition_operator(left, str(operator), right)
+        return apply_condition_operator(left, str(operator), right, value_kind)
 
     async def _loop(self, node: Dict, nodeOutputs: Dict, nodeId: str, inputSources: Dict) -> Any:
         params = node.get("parameters") or {}
         itemsPath = params.get("items", "[]")
         from modules.workflows.automation2.graphUtils import resolveParameterReferences
 
-        raw = resolveParameterReferences(itemsPath, nodeOutputs)
+        raw = resolveParameterReferences(
+            itemsPath,
+            nodeOutputs,
+            consumer_node_id=nodeId,
+            input_sources=inputSources,
+        )
         items = self._normalize_loop_items(raw)
         mode = (params.get("iterationMode") or "all").strip().lower()
         stride = params.get("iterationStride", 2)
@@ -245,6 +278,8 @@ class FlowExecutor:
 
     def _normalize_loop_items(self, raw: Any) -> List[Any]:
         """Coerce resolved `items` into a list (lists, dict children, or scalars)."""
+        if isinstance(raw, dict) and isinstance(raw.get("items"), list):
+            return self._expand_presentation_lines_loop_items(raw["items"])
         if isinstance(raw, list):
             return self._expand_presentation_lines_loop_items(raw)
         if isinstance(raw, dict):

modules/workflows/automation2/graphUtils.py

@@ -253,6 +253,8 @@ def _checkPortCompatibility(
                 continue
             srcOutputPorts = srcDef.get("outputPorts", {})
             srcPort = srcOutputPorts.get(srcOut, {}) or {}
+            if srcNode.get("type") == "flow.switch" and not srcPort.get("schema"):
+                srcPort = srcOutputPorts.get(0, {}) or srcPort
             tgtPort = tgtInputPorts.get(tgtIn, {}) or {}
 
             if not isinstance(srcPort, dict):
@@ -264,6 +266,9 @@ def _checkPortCompatibility(
                 continue
             if src_schema in accepts:
                 continue
+            # ContextBranch is a typed Transit envelope (switch filtered branches).
+            if src_schema == "ContextBranch" and ("Transit" in accepts or "ContextBranch" in accepts):
+                continue
             # Port that only declares Transit behaves as an untyped sink (legacy graphs).
             if len(accepts) == 1 and accepts[0] == "Transit":
                 continue
@@ -409,12 +414,21 @@ def _unwrapTypedRef(value: Any) -> Any:
     return value.get(primary, value)
 
 
-def resolveParameterReferences(value: Any, nodeOutputs: Dict[str, Any]) -> Any:
+def resolveParameterReferences(
+    value: Any,
+    nodeOutputs: Dict[str, Any],
+    *,
+    consumer_node_id: Optional[str] = None,
+    input_sources: Optional[Dict[str, Dict[int, tuple]]] = None,
+) -> Any:
     """
     Resolve parameter references:
     - {{nodeId.output}} or {{nodeId.output.path}} in strings (legacy)
     - { "type": "ref", "nodeId": "...", "path": ["field", "nested"] } -> resolved value
     - { "type": "value", "value": ... } -> value (then recursively resolve)
+
+    When ``consumer_node_id`` and ``input_sources`` are set, refs to the wired
+    upstream switch use that connection's output port (per-branch payload).
     """
     import json
     import re
@@ -430,8 +444,13 @@ def resolveParameterReferences(value: Any, nodeOutputs: Dict[str, Any]) -> Any:
             path = value.get("path")
             if node_id is not None and isinstance(path, (list, tuple)):
                 data = nodeOutputs.get(node_id)
-                # Unwrap transit envelopes to access the real data
-                if isinstance(data, dict) and data.get("_transit"):
+                wired = None
+                if consumer_node_id and input_sources:
+                    wired = (input_sources.get(consumer_node_id) or {}).get(0)
+                if wired and wired[0] == node_id:
+                    from modules.features.graphicalEditor.switchOutput import unwrap_transit_for_port
+                    data = unwrap_transit_for_port(data, wired[1])
+                elif isinstance(data, dict) and data.get("_transit"):
                     data = data.get("data", data)
                 plist = list(path)
                 resolved = _get_by_path(data, plist)
@@ -450,16 +469,34 @@ def resolveParameterReferences(value: Any, nodeOutputs: Dict[str, Any]) -> Any:
                         # Form nodes store fields under {"payload": {fieldName: …}}.
                         # DataPicker emits bare field paths like ["url"]; try under payload.
                         resolved = _get_by_path(data["payload"], plist)
-                return resolveParameterReferences(resolved, nodeOutputs)
+                return resolveParameterReferences(
+                    resolved,
+                    nodeOutputs,
+                    consumer_node_id=consumer_node_id,
+                    input_sources=input_sources,
+                )
             return value
         if value.get("type") == "value":
             inner = value.get("value")
-            return resolveParameterReferences(inner, nodeOutputs)
+            return resolveParameterReferences(
+                inner,
+                nodeOutputs,
+                consumer_node_id=consumer_node_id,
+                input_sources=input_sources,
+            )
         if value.get("type") == "system":
             variable = value.get("variable", "")
             from modules.features.graphicalEditor.portTypes import resolveSystemVariable
             return resolveSystemVariable(variable, nodeOutputs.get("_context", {}))
-        return {k: resolveParameterReferences(v, nodeOutputs) for k, v in value.items()}
+        return {
+            k: resolveParameterReferences(
+                v,
+                nodeOutputs,
+                consumer_node_id=consumer_node_id,
+                input_sources=input_sources,
+            )
+            for k, v in value.items()
+        }
 
     if isinstance(value, str):
         def repl(m):
@@ -498,11 +535,27 @@ def resolveParameterReferences(value: Any, nodeOutputs: Dict[str, Any]) -> Any:
         # contextBuilder: list where every item is a `{"type":"ref",...}` envelope.
         # Resolve each part; a single ref preserves the resolved type (str, list, dict).
         if value and all(isinstance(v, dict) and v.get("type") == "ref" for v in value):
-            resolved_parts = [resolveParameterReferences(v, nodeOutputs) for v in value]
+            resolved_parts = [
+                resolveParameterReferences(
+                    v,
+                    nodeOutputs,
+                    consumer_node_id=consumer_node_id,
+                    input_sources=input_sources,
+                )
+                for v in value
+            ]
             if len(resolved_parts) == 1:
                 return resolved_parts[0]
             return resolved_parts
-        return [resolveParameterReferences(v, nodeOutputs) for v in value]
+        return [
+            resolveParameterReferences(
+                v,
+                nodeOutputs,
+                consumer_node_id=consumer_node_id,
+                input_sources=input_sources,
+            )
+            for v in value
+        ]
     return value
 
 

modules/workflows/methods/methodContext/actions/extractContent.py

@@ -934,6 +934,52 @@ def _presentation_image_marker_in_data(part: Dict[str, Any]) -> Dict[str, Any]:
     return marker
 
 
+_BLOB_IMAGE_CHUNK_RE = re.compile(r"^\[image(?:\:([^\]]+))?\]$")
+
+
+def parse_blob_data_segments(data: str) -> List[Dict[str, Any]]:
+    """Split presentation ``blob`` ``data`` into virtual slots (text chunks + image markers)."""
+    segments: List[Dict[str, Any]] = []
+    if not isinstance(data, str) or not data.strip():
+        return segments
+    for idx, chunk in enumerate(data.split("\n\n")):
+        piece = chunk.strip()
+        if not piece:
+            continue
+        m = _BLOB_IMAGE_CHUNK_RE.fullmatch(piece)
+        if m:
+            token = (m.group(1) or "").strip()
+            seg: Dict[str, Any] = {"typeGroup": "image", "mimeType": "image/*", "data": piece}
+            if token:
+                seg["id"] = token
+            else:
+                seg["id"] = f"blob_image_{idx}"
+            segments.append(seg)
+        else:
+            segments.append({"typeGroup": "text", "mimeType": "text/plain", "data": piece, "id": f"blob_text_{idx}"})
+    return segments
+
+
+def filter_blob_bucket_by_content_type(bucket: Dict[str, Any], content_type: str) -> Dict[str, Any]:
+    """Keep only blob segments matching ``content_type`` (re-join as ``\\n\\n`` string)."""
+    out = copy.deepcopy(bucket)
+    raw = out.get("data")
+    if not isinstance(raw, str):
+        return out
+    target = (content_type or "").strip().lower()
+    kept: List[str] = []
+    for seg in parse_blob_data_segments(raw):
+        tg = (seg.get("typeGroup") or "").strip().lower()
+        if target == "media" and tg in ("image", "media", "video", "audio"):
+            kept.append(str(seg.get("data") or ""))
+        elif target == "text" and tg in ("text", "table", "structure"):
+            kept.append(str(seg.get("data") or ""))
+        elif tg == target:
+            kept.append(str(seg.get("data") or ""))
+    out["data"] = "\n\n".join(s for s in kept if s.strip())
+    return out
+
+
 def _build_file_presentation(
     source_file_name: str,
     parts: List[Dict[str, Any]],
@@ -959,8 +1005,8 @@ def _build_file_presentation(
             tg = (p.get("typeGroup") or "").strip()
             if tg == "image":
                 m = _presentation_image_marker_in_data(p)
-                pid = str(m.get("partId") or "").strip()
-                chunks_blob.append(f"[image:{pid}]" if pid else "[image]")
+                token = str(m.get("embeddedImageFileId") or m.get("partId") or "").strip()
+                chunks_blob.append(f"[image:{token}]" if token else "[image]")
                 continue
             if _part_carries_plain_text(p):
                 raw = p.get("data")
@@ -1433,6 +1479,20 @@ def normalize_presentation_envelopes(raw: Any) -> List[Dict[str, Any]]:
                     file_key=str(raw.get("name") or "file_1"),
                 )
             ]
+        if isinstance(raw.get("name"), str) and isinstance(raw.get("value"), dict):
+            slot = raw["value"]
+            if _is_presentation_line_slot(slot):
+                bucket = {
+                    "outputMode": slot.get("outputMode") or "lines",
+                    "sourceFileName": "",
+                    "data": [slot],
+                }
+                return [
+                    presentation_envelope_from_file_bucket(
+                        bucket,
+                        file_key=str(raw.get("name") or "file_1"),
+                    )
+                ]
         if _is_presentation_file_bucket(raw):
             return [presentation_envelope_from_file_bucket(raw)]
         if _is_presentation_line_slot(raw):
@@ -1450,6 +1510,27 @@ def normalize_presentation_envelopes(raw: Any) -> List[Dict[str, Any]]:
     return []
 
 
+def _artifacts_by_part_id_from_meta(meta: Any) -> Dict[str, str]:
+    out: Dict[str, str] = {}
+    if not isinstance(meta, dict):
+        return out
+    for art in meta.get("persistedImageArtifacts") or []:
+        if not isinstance(art, dict):
+            continue
+        sp = str(art.get("sourcePartId") or "").strip()
+        fid = str(art.get("fileId") or "").strip()
+        if sp and fid:
+            out[sp] = fid
+    return out
+
+
+def _collect_artifacts_by_part_id(envelopes: List[Dict[str, Any]]) -> Dict[str, str]:
+    merged: Dict[str, str] = {}
+    for envelope in envelopes:
+        merged.update(_artifacts_by_part_id_from_meta(envelope.get("_meta")))
+    return merged
+
+
 def presentation_envelopes_to_document_json(
     raw: Any,
     *,
@@ -1466,6 +1547,8 @@ def presentation_envelopes_to_document_json(
             "context must be presentation data from Inhalt extrahieren (kind=context.extractContent.presentation.v1)"
         )
 
+    artifacts_by_part = _collect_artifacts_by_part_id(envelopes)
+
     sections: List[Dict[str, Any]] = []
     order = 0
 
@@ -1496,8 +1579,35 @@ def presentation_envelopes_to_document_json(
             "elements": [{"content": {"inlineRuns": _parseInlineRuns(t)}}],
         })
 
-    def _append_image_slot(slot: Dict[str, Any]) -> None:
+    def _resolve_image_file_id(slot: Dict[str, Any]) -> Optional[str]:
         fid = slot.get("embeddedImageFileId")
+        if fid:
+            return str(fid).strip() or None
+        candidates: List[str] = []
+        sid = str(slot.get("id") or "").strip()
+        if sid:
+            candidates.append(sid)
+        raw_d = slot.get("data")
+        if isinstance(raw_d, str):
+            m = _BLOB_IMAGE_CHUNK_RE.fullmatch(raw_d.strip())
+            if m:
+                tok = (m.group(1) or "").strip()
+                if tok:
+                    candidates.append(tok)
+        for cand in candidates:
+            if cand in artifacts_by_part:
+                return artifacts_by_part[cand]
+            # Marker may already carry the persisted storage file id.
+            try:
+                blob = _load_image_bytes_by_file_id(services, cand)
+                if blob:
+                    return cand
+            except Exception:
+                pass
+        return None
+
+    def _append_image_slot(slot: Dict[str, Any]) -> None:
+        fid = _resolve_image_file_id(slot)
         if not fid:
             raise ValueError(
                 "image slot is missing embeddedImageFileId — "
@@ -1589,6 +1699,11 @@ def presentation_envelopes_to_document_json(
             if src:
                 _append_heading(src)
         raw_data = bucket.get("data")
+        mode = str(bucket.get("outputMode") or "").strip().lower()
+        if isinstance(raw_data, str) and mode == "blob":
+            for seg in parse_blob_data_segments(raw_data):
+                _append_slot(seg)
+            return
         if isinstance(raw_data, str):
             _append_paragraph(raw_data)
             return

tests/unit/graphicalEditor/test_upstream_paths_and_graph_schema.py

@@ -50,6 +50,25 @@ def test_parse_graph_defined_schema_nested_group():
     assert "addr.zip" in names
 
 
+def test_compute_upstream_paths_switch_context_branch_items():
+    graph = {
+        "nodes": [
+            {"id": "ext1", "type": "context.extractContent", "parameters": {}},
+            {"id": "sw1", "type": "flow.switch", "parameters": {"cases": [{"operator": "contains_content", "value": "image"}]}},
+            {"id": "ai1", "type": "ai.prompt", "parameters": {"aiPrompt": "summarize"}},
+        ],
+        "connections": [
+            {"source": "ext1", "target": "sw1", "sourceOutput": 0, "targetInput": 0},
+            {"source": "sw1", "target": "ai1", "sourceOutput": 0, "targetInput": 0},
+        ],
+    }
+    paths = compute_upstream_paths(graph, "ai1")
+    sw_paths = [p for p in paths if p.get("producerNodeId") == "sw1"]
+    items_paths = [p for p in sw_paths if p.get("path") == ["items"]]
+    assert items_paths, sw_paths
+    assert items_paths[0].get("type") == "List[Any]"
+
+
 def test_validate_graph_port_mismatch_errors():
     node_type_ids = {n["id"] for n in STATIC_NODE_TYPES}
     graph = {

tests/unit/workflow/test_flow_executor_conditions.py

@@ -61,6 +61,34 @@ def test_context_contains_content(executor):
     assert executor._evalStructuredCondition(cond, {"n1": presentation}, item_param={"type": "ref", "nodeId": "n1", "path": []}, node={"id": "if1", "type": "flow.ifElse"}) is True
 
 
+def test_context_contains_content_blob_mode(executor):
+    presentation = {
+        "kind": PRESENTATION_KIND,
+        "outputMode": "blob",
+        "fileOrder": ["f1"],
+        "files": {
+            "f1": {
+                "outputMode": "blob",
+                "data": "Invoice text\n\n[image:abc123]",
+            }
+        },
+    }
+    img_cond = {"type": "condition", "operator": "contains_content", "value": "image"}
+    txt_cond = {"type": "condition", "operator": "contains_content", "value": "text"}
+    item = {"type": "ref", "nodeId": "n1", "path": []}
+    node = {"id": "if1", "type": "flow.ifElse"}
+    assert executor._evalStructuredCondition(img_cond, {"n1": presentation}, item_param=item, node=node) is True
+    assert executor._evalStructuredCondition(txt_cond, {"n1": presentation}, item_param=item, node=node) is True
+
+
 def test_switch_uses_shared_operators(executor):
     assert executor._evalSwitchCase("abc", {"operator": "starts_with", "value": "ab"}) is True
     assert executor._evalSwitchCase([1, 2], {"operator": "length_eq", "value": 2}) is True
+
+
+def test_switch_resolves_value_kind_for_string_ops(executor):
+    assert executor._evalSwitchCase(
+        "hello",
+        {"operator": "starts_with", "value": "he"},
+        value_kind="string",
+    ) is True

tests/unit/workflow/test_switch_filtered_output.py

@@ -0,0 +1,359 @@
+# Copyright (c) 2025 Patrick Motsch
+"""flow.switch ContextBranch: filtered presentation + loop-ready items."""
+
+import pytest
+
+from modules.features.graphicalEditor.portTypes import unwrapTransit, wrapTransit
+from modules.features.graphicalEditor.switchOutput import (
+    build_switch_branch_payload,
+    build_switch_combined_output,
+    build_switch_default_payload,
+    unwrap_transit_for_port,
+)
+from modules.workflows.automation2.executionEngine import _is_node_on_active_path
+from modules.workflows.automation2.executors.flowExecutor import FlowExecutor
+from modules.workflows.automation2.graphUtils import resolveParameterReferences
+from modules.workflows.methods.methodContext.actions.extractContent import PRESENTATION_KIND
+
+
+def _presentation_with_text_and_image():
+    return {
+        "kind": PRESENTATION_KIND,
+        "schemaVersion": "1",
+        "outputMode": "parts",
+        "fileOrder": ["doc"],
+        "files": {
+            "doc": {
+                "outputMode": "parts",
+                "data": [
+                    {"typeGroup": "text", "id": "t1", "data": "Hello"},
+                    {"typeGroup": "image", "id": "i1", "mimeType": "image/png", "data": "YQ=="},
+                ],
+            }
+        },
+    }
+
+
+def _presentation_blob_with_text_and_image():
+    blob_data = "Hello world\n\n[image:img1]\n\nMore text"
+    return {
+        "kind": PRESENTATION_KIND,
+        "schemaVersion": "1",
+        "outputMode": "blob",
+        "fileOrder": ["doc"],
+        "files": {
+            "doc": {
+                "outputMode": "blob",
+                "sourceFileName": "test.pdf",
+                "data": blob_data,
+            }
+        },
+    }
+
+
+def test_build_switch_branch_payload_filters_blob_image():
+    pres = _presentation_blob_with_text_and_image()
+    payload = build_switch_branch_payload(
+        pres,
+        {"operator": "contains_content", "value": "image"},
+        value_kind="context",
+        match_index=0,
+    )
+    assert payload["filterApplied"] is True
+    assert len(payload["items"]) == 1
+    assert payload["items"][0]["value"]["typeGroup"] == "image"
+    assert "[image:img1]" in payload["data"]["files"]["doc"]["data"]
+
+
+def test_build_switch_branch_payload_filters_blob_text():
+    pres = _presentation_blob_with_text_and_image()
+    payload = build_switch_branch_payload(
+        pres,
+        {"operator": "contains_content", "value": "text"},
+        value_kind="context",
+        match_index=1,
+    )
+    assert payload["filterApplied"] is True
+    assert len(payload["items"]) == 2
+    assert all(i["value"]["typeGroup"] == "text" for i in payload["items"])
+    filtered = payload["data"]["files"]["doc"]["data"]
+    assert "Hello world" in filtered
+    assert "[image:" not in filtered
+
+
+@pytest.mark.asyncio
+async def test_switch_blob_multi_match():
+    executor = FlowExecutor()
+    pres = _presentation_blob_with_text_and_image()
+    sw_id = "sw1"
+    node_outputs = {
+        "ext1": pres,
+        "_context": {
+            "graphNodesById": {
+                "ext1": {"id": "ext1", "type": "context.extractContent"},
+                sw_id: {"id": sw_id, "type": "flow.switch"},
+            }
+        },
+    }
+    executor._getInputData = lambda *_a, **_k: pres  # type: ignore[method-assign]
+    node = {
+        "id": sw_id,
+        "type": "flow.switch",
+        "parameters": {
+            "value": {"type": "ref", "nodeId": "ext1", "path": []},
+            "cases": [
+                {"operator": "contains_content", "value": "image"},
+                {"operator": "contains_content", "value": "text"},
+            ],
+        },
+    }
+    out = await executor._switch(node, node_outputs, sw_id, {})
+    assert out["_meta"]["matches"] == [0, 1]
+    assert len(unwrap_transit_for_port(out, 0)["items"]) == 1
+    assert len(unwrap_transit_for_port(out, 1)["items"]) == 2
+
+
+def test_switch_blob_image_items_get_embedded_file_id():
+    part_id = "dbd27119-cd21-4a62-b5e2-b06d3b81470b"
+    file_id = "storage-file-uuid-1"
+    pres = {
+        "kind": PRESENTATION_KIND,
+        "outputMode": "blob",
+        "fileOrder": ["doc"],
+        "files": {
+            "doc": {
+                "outputMode": "blob",
+                "data": f"Hello\n\n[image:{part_id}]",
+            }
+        },
+        "_meta": {
+            "persistedImageArtifacts": [
+                {"sourcePartId": part_id, "fileId": file_id, "mimeType": "image/png"},
+            ]
+        },
+    }
+    payload = build_switch_branch_payload(
+        pres,
+        {"operator": "contains_content", "value": "image"},
+        value_kind="context",
+        match_index=0,
+    )
+    assert len(payload["items"]) == 1
+    slot = payload["items"][0]["value"]
+    assert slot.get("embeddedImageFileId") == file_id
+
+
+def test_build_switch_branch_payload_filters_images():
+    pres = _presentation_with_text_and_image()
+    case = {"operator": "contains_content", "value": "image"}
+    payload = build_switch_branch_payload(
+        pres,
+        case,
+        value_kind="context",
+        match_index=0,
+    )
+    assert payload["filterApplied"] is True
+    assert payload["contentType"] == "image"
+    assert len(payload["items"]) == 1
+    assert payload["items"][0]["value"]["typeGroup"] == "image"
+    data = payload["data"]
+    assert data["kind"] == PRESENTATION_KIND
+    slots = data["files"]["doc"]["data"]
+    assert len(slots) == 1
+    assert slots[0]["typeGroup"] == "image"
+
+
+def test_build_switch_default_payload_passthrough():
+    pres = _presentation_with_text_and_image()
+    payload = build_switch_default_payload(pres, match_index=2)
+    assert payload["filterApplied"] is False
+    assert payload["match"] == 2
+    assert payload["data"]["fileOrder"] == pres["fileOrder"]
+    assert len(payload["items"]) == 2
+
+
+@pytest.mark.asyncio
+async def test_switch_executor_match_and_default_branch():
+    executor = FlowExecutor()
+    pres = _presentation_with_text_and_image()
+    ext_id = "ext1"
+    sw_id = "sw1"
+    node_outputs = {
+        ext_id: pres,
+        "_context": {
+            "graphNodesById": {
+                ext_id: {"id": ext_id, "type": "context.extractContent"},
+                sw_id: {"id": sw_id, "type": "flow.switch"},
+            }
+        },
+    }
+
+    def _inp(_nid, _sources, _outputs, _output_index=0):
+        return pres
+
+    executor._getInputData = _inp  # type: ignore[method-assign]
+
+    match_node = {
+        "id": sw_id,
+        "type": "flow.switch",
+        "parameters": {
+            "value": {"type": "ref", "nodeId": ext_id, "path": []},
+            "cases": [{"operator": "contains_content", "value": "image"}],
+        },
+    }
+    match_out = await executor._switch(match_node, node_outputs, sw_id, {})
+    match_payload = unwrapTransit(match_out)
+    assert match_out["_meta"]["match"] == 0
+    assert match_out["_meta"]["matches"] == [0]
+    assert match_payload["filterApplied"] is True
+    assert len(match_payload["items"]) == 1
+    assert match_payload["branches"]["0"]["contentType"] == "image"
+
+    default_node = {
+        **match_node,
+        "parameters": {
+            **match_node["parameters"],
+            "cases": [{"operator": "contains_content", "value": "video"}],
+        },
+    }
+    default_out = await executor._switch(default_node, node_outputs, sw_id, {})
+    assert default_out["_meta"]["match"] == 1
+    assert default_out["_meta"]["matches"] == [1]
+    default_payload = unwrapTransit(default_out)
+    assert default_payload["filterApplied"] is False
+    assert default_payload["data"]["fileOrder"] == pres["fileOrder"]
+
+
+@pytest.mark.asyncio
+async def test_switch_multi_match_text_and_image_branches():
+    executor = FlowExecutor()
+    pres = _presentation_with_text_and_image()
+    sw_id = "sw1"
+    node_outputs = {
+        "ext1": pres,
+        "_context": {
+            "graphNodesById": {
+                "ext1": {"id": "ext1", "type": "context.extractContent"},
+                sw_id: {"id": sw_id, "type": "flow.switch"},
+            }
+        },
+    }
+    executor._getInputData = lambda *_a, **_k: pres  # type: ignore[method-assign]
+
+    node = {
+        "id": sw_id,
+        "type": "flow.switch",
+        "parameters": {
+            "value": {"type": "ref", "nodeId": "ext1", "path": []},
+            "cases": [
+                {"operator": "contains_content", "value": "image"},
+                {"operator": "contains_content", "value": "text"},
+            ],
+        },
+    }
+    out = await executor._switch(node, node_outputs, sw_id, {})
+    assert out["_meta"]["matches"] == [0, 1]
+    img = unwrap_transit_for_port(out, 0)
+    txt = unwrap_transit_for_port(out, 1)
+    assert img["contentType"] == "image"
+    assert txt["contentType"] == "text"
+    assert len(img["items"]) == 1
+    assert len(txt["items"]) == 1
+    assert img["items"][0]["value"]["typeGroup"] == "image"
+    assert txt["items"][0]["value"]["typeGroup"] == "text"
+
+
+def test_active_path_allows_all_matching_switch_ports():
+    combined = build_switch_combined_output(
+        _presentation_with_text_and_image(),
+        [
+            {"operator": "contains_content", "value": "image"},
+            {"operator": "contains_content", "value": "text"},
+        ],
+        matched_indices=[0, 1],
+        value_kind="context",
+    )
+    sw_out = wrapTransit(combined, {"match": 0, "matches": [0, 1]})
+    node_outputs = {"sw1": sw_out}
+    conn_map = {
+        "loop_img": [("sw1", 0, 0)],
+        "file_txt": [("sw1", 1, 0)],
+    }
+    assert _is_node_on_active_path("loop_img", conn_map, node_outputs)
+    assert _is_node_on_active_path("file_txt", conn_map, node_outputs)
+    assert not _is_node_on_active_path("other", {"other": [("sw1", 2, 0)]}, node_outputs)
+
+
+@pytest.mark.asyncio
+async def test_loop_uses_switch_items_ref():
+    executor = FlowExecutor()
+    pres = _presentation_with_text_and_image()
+    branch = build_switch_branch_payload(
+        pres,
+        {"operator": "contains_content", "value": "image"},
+        value_kind="context",
+        match_index=0,
+    )
+    sw_id = "sw1"
+    node_outputs = {sw_id: wrapTransit(branch, {"match": 0})}
+
+    loop_node = {
+        "id": "loop1",
+        "type": "flow.loop",
+        "parameters": {
+            "items": {"type": "ref", "nodeId": sw_id, "path": ["items"]},
+        },
+    }
+    out = await executor._loop(loop_node, node_outputs, "loop1", {})
+    assert out["count"] == 1
+    assert out["items"][0]["value"]["typeGroup"] == "image"
+
+
+def test_resolve_context_builder_ref_uses_switch_output_port():
+    """file.create context ref to switch.items must use the wired source output port."""
+    pres = _presentation_with_text_and_image()
+    combined = build_switch_combined_output(
+        pres,
+        [
+            {"operator": "contains_content", "value": "image"},
+            {"operator": "contains_content", "value": "text"},
+        ],
+        matched_indices=[0, 1],
+        value_kind="context",
+    )
+    sw_id = "sw1"
+    consumer_id = "fc1"
+    node_outputs = {sw_id: wrapTransit(combined, {"match": 0, "matches": [0, 1]})}
+    input_sources = {consumer_id: {0: (sw_id, 1)}}
+    resolved = resolveParameterReferences(
+        {
+            "context": [
+                {
+                    "type": "ref",
+                    "nodeId": sw_id,
+                    "path": ["items"],
+                }
+            ],
+        },
+        node_outputs,
+        consumer_node_id=consumer_id,
+        input_sources=input_sources,
+    )
+    items = resolved["context"]
+    assert isinstance(items, list)
+    assert len(items) == 1
+    assert items[0]["value"]["typeGroup"] == "text"
+    branch = build_switch_branch_payload(
+        _presentation_with_text_and_image(),
+        {"operator": "contains_content", "value": "image"},
+        value_kind="context",
+        match_index=0,
+    )
+    node_outputs = {"sw1": wrapTransit(branch, {"match": 0})}
+    resolved = resolveParameterReferences(
+        {"type": "ref", "nodeId": "sw1", "path": ["items"]},
+        node_outputs,
+    )
+    assert isinstance(resolved, list)
+    assert len(resolved) == 1
+    assert resolved[0]["value"]["typeGroup"] == "image"

tests/unit/workflows/test_automation2_graphUtils.py

@@ -37,6 +37,34 @@ class TestValidateGraphStartNode:
         assert not any("no start node" in e.lower() for e in errs)
 
 
+    def test_switch_second_output_to_ai_prompt_ok(self):
+        from modules.features.graphicalEditor.nodeDefinitions import STATIC_NODE_TYPES
+
+        node_type_ids = {n["id"] for n in STATIC_NODE_TYPES}
+        graph = {
+            "nodes": [
+                {"id": "t", "type": "trigger.manual", "parameters": {}},
+                {
+                    "id": "sw",
+                    "type": "flow.switch",
+                    "parameters": {
+                        "cases": [
+                            {"operator": "contains_content", "value": "image"},
+                            {"operator": "contains_content", "value": "text"},
+                        ],
+                    },
+                },
+                {"id": "ai", "type": "ai.prompt", "parameters": {"aiPrompt": "hi"}},
+            ],
+            "connections": [
+                {"source": "sw", "target": "ai", "sourceOutput": 1, "targetInput": 0},
+            ],
+        }
+        errs = validateGraph(graph, node_type_ids)
+        port_errs = [e for e in errs if "Port mismatch" in e]
+        assert port_errs == [], port_errs
+
+
 class TestResolveParameterReferences:
     """Test structured ref/value resolution."""