execweave 0.8.5__tar.gz → 0.8.6__tar.gz

{execweave-0.8.5 → execweave-0.8.6}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.5
 Name: execweave
-Version: 0.8.5
+Version: 0.8.6
 Summary: Turn AI-agent runtime behavior into a local interactive execution graph.
 Project-URL: Homepage, https://github.com/Irish-kw/ExecWeave
 Project-URL: Repository, https://github.com/Irish-kw/ExecWeave
@@ -83,7 +83,7 @@ ExecWeave is a source-available, local-first observability project that turns AI
   <img src="docs/assets/codex.gif" alt="ExecWeave animated live demo" width="100%">
 </p>
 
-This README documents **v0.8.5**.
+This README documents **v0.8.6**.
 
 ## Why ExecWeave
 

{execweave-0.8.5 → execweave-0.8.6}/README.de.md

@@ -30,7 +30,7 @@ ExecWeave ist ein source-available, local-first Observability-Projekt, das Aktiv
   <img src="docs/assets/codex.gif" alt="ExecWeave animated live demo" width="100%">
 </p>
 
-Dieses README dokumentiert **v0.8.5**.
+Dieses README dokumentiert **v0.8.6**.
 
 ## Warum ExecWeave
 

{execweave-0.8.5 → execweave-0.8.6}/README.fr.md

@@ -30,7 +30,7 @@ ExecWeave est un projet d’observabilité source-available et local-first qui t
   <img src="docs/assets/codex.gif" alt="ExecWeave animated live demo" width="100%">
 </p>
 
-Ce README décrit **v0.8.5**.
+Ce README décrit **v0.8.6**.
 
 ## Pourquoi ExecWeave
 

{execweave-0.8.5 → execweave-0.8.6}/README.ja.md

@@ -30,7 +30,7 @@ ExecWeave は source-available / local-first の observability プロジェク
   <img src="docs/assets/codex.gif" alt="ExecWeave animated live demo" width="100%">
 </p>
 
-この README は **v0.8.5** を説明します。
+この README は **v0.8.6** を説明します。
 
 ## ExecWeave を使う理由
 

{execweave-0.8.5 → execweave-0.8.6}/README.ko.md

@@ -30,7 +30,7 @@ ExecWeave는 source-available, local-first observability 프로젝트입니다.
   <img src="docs/assets/codex.gif" alt="ExecWeave animated live demo" width="100%">
 </p>
 
-이 README는 **v0.8.5**을 설명합니다.
+이 README는 **v0.8.6**을 설명합니다.
 
 ## 왜 ExecWeave인가
 

{execweave-0.8.5 → execweave-0.8.6}/README.md

@@ -30,7 +30,7 @@ ExecWeave is a source-available, local-first observability project that turns AI
   <img src="docs/assets/codex.gif" alt="ExecWeave animated live demo" width="100%">
 </p>
 
-This README documents **v0.8.5**.
+This README documents **v0.8.6**.
 
 ## Why ExecWeave
 

{execweave-0.8.5 → execweave-0.8.6}/README.ru.md

@@ -30,7 +30,7 @@ ExecWeave — source-available, local-first проект наблюдаемос
   <img src="docs/assets/codex.gif" alt="ExecWeave animated live demo" width="100%">
 </p>
 
-Этот README описывает **v0.8.5**.
+Этот README описывает **v0.8.6**.
 
 ## Почему ExecWeave
 

{execweave-0.8.5 → execweave-0.8.6}/README.zh-CN.md

@@ -30,7 +30,7 @@ ExecWeave 是一个 source-available、local-first 的可观测性项目，把 A
   <img src="docs/assets/codex.gif" alt="ExecWeave animated live demo" width="100%">
 </p>
 
-本 README 对应 **v0.8.5**。
+本 README 对应 **v0.8.6**。
 
 ## 为什么是 ExecWeave
 

{execweave-0.8.5 → execweave-0.8.6}/README.zh-TW.md

@@ -30,7 +30,7 @@ ExecWeave 是一個 source-available、local-first 的可觀測性專案，會
   <img src="docs/assets/codex.gif" alt="ExecWeave animated live demo" width="100%">
 </p>
 
-本 README 對應 **v0.8.5**。
+本 README 對應 **v0.8.6**。
 
 ## 為什麼是 ExecWeave
 

{execweave-0.8.5 → execweave-0.8.6}/pyproject.toml

@@ -4,8 +4,8 @@ build-backend = "hatchling.build"
 
 [project]
 name = "execweave"
-# ExecWeave v0.8.5 release metadata
-version = "0.8.5"
+# ExecWeave v0.8.6 release metadata
+version = "0.8.6"
 description = "Turn AI-agent runtime behavior into a local interactive execution graph."
 readme = "README.md"
 requires-python = ">=3.10"

{execweave-0.8.5 → execweave-0.8.6}/src/execweave/__init__.py

@@ -1,3 +1,3 @@
 """ExecWeave runtime graph package."""
 
-__version__ = "0.8.5"
+__version__ = "0.8.6"

{execweave-0.8.5 → execweave-0.8.6}/src/execweave/antigravity_subagent_linkage.py

@@ -115,7 +115,9 @@ def _request_specs(value: object) -> list[dict[str, Any]] | None:
     if not isinstance(value, list) or not value:
         return None
     specs: list[dict[str, Any]] = []
-    allowed = {"Prompt", "Role", "TypeName", "Workspace"}
+    # Current Antigravity includes Model in each invoke_subagent spec.
+    # Keep the allow-list strict while accepting that live-verified field.
+    allowed = {"Model", "Prompt", "Role", "TypeName", "Workspace"}
     for raw in value:
         if not isinstance(raw, dict) or not set(raw).issubset(allowed):
             return None
@@ -128,6 +130,9 @@ def _request_specs(value: object) -> list[dict[str, Any]] | None:
             return None
         if not isinstance(type_name, str) or not type_name:
             return None
+        model = raw.get("Model")
+        if model is not None and (not isinstance(model, str) or not model):
+            return None
         workspace = raw.get("Workspace")
         if workspace is not None and (not isinstance(workspace, str) or not workspace):
             return None
@@ -211,9 +216,12 @@ def _matching_request(record: dict[str, Any], specs: list[dict[str, Any]]) -> bo
 
 
 def _matching_result(record: dict[str, Any]) -> list[dict[str, Any]] | None:
-    if record.get("source") != "MODEL":
+    if record.get("source") != "MODEL" or record.get("status") != "DONE":
         return None
-    if record.get("type") != "INVOKE_SUBAGENT" or record.get("status") != "DONE":
+    # Older builds labelled the immediate result INVOKE_SUBAGENT; the current
+    # live wire labels that same record GENERIC. GENERIC is accepted only if its
+    # payload passes _parse_result_content's exact provider prefix/schema checks.
+    if record.get("type") not in {"INVOKE_SUBAGENT", "GENERIC"}:
         return None
     return _parse_result_content(record.get("content"))
 

{execweave-0.8.5 → execweave-0.8.6}/src/execweave/conversation_records.py

@@ -235,6 +235,123 @@ def _history_message_key(message: dict[str, Any]) -> tuple[object, ...]:
     )
 
 
+def _project_antigravity_addressed_tasks(
+    entries: list[dict[str, Any]],
+    graph: dict[str, Any],
+) -> None:
+    """Project exact parent-addressed send_message text into the child timeline.
+
+    Raw Antigravity topology already carries positive ``parent_scope_id`` evidence on
+    each validated child node. Raw send_message conversation evidence carries exact
+    provider sender and recipient conversation IDs. Join those two facts only for
+    presentation: an addressed parent message becomes a child task opener, while raw
+    evidence remains unchanged and delivery/consumption remain explicitly unobserved.
+    """
+    prefix = "agent:antigravity:conversation:"
+    topology: dict[str, tuple[str, str]] = {}
+    for node in graph.get("nodes", []):
+        if not isinstance(node, dict) or node.get("type") != "agent":
+            continue
+        node_id = node.get("id")
+        attrs = node.get("attributes")
+        attrs = attrs if isinstance(attrs, dict) else {}
+        if not isinstance(node_id, str) or not node_id.startswith(prefix):
+            continue
+        if str(attrs.get("provider") or "").lower() != "antigravity":
+            continue
+        parent_path = attrs.get("parent_agent_path")
+        parent_scope = attrs.get("parent_scope_id")
+        if not isinstance(parent_path, str) or not parent_path:
+            continue
+        if not isinstance(parent_scope, str) or not parent_scope:
+            continue
+        topology[node_id.removeprefix(prefix)] = (parent_scope, parent_path)
+
+    children: dict[str, dict[str, Any]] = {}
+    for entry in entries:
+        if str(entry.get("provider") or "").lower() != "antigravity":
+            continue
+        source_id = entry.get("source_id")
+        preview = entry.get("conversation_preview")
+        if not isinstance(source_id, str) or not source_id.startswith(prefix):
+            continue
+        child_id = source_id.removeprefix(prefix)
+        if child_id not in topology or not isinstance(preview, dict):
+            continue
+        children[child_id] = entry
+
+    additions: dict[str, list[dict[str, Any]]] = {child_id: [] for child_id in children}
+    for entry in entries:
+        if str(entry.get("provider") or "").lower() != "antigravity":
+            continue
+        preview = entry.get("conversation_preview")
+        if not isinstance(preview, dict):
+            continue
+        for message in preview.get("messages") or []:
+            if not isinstance(message, dict) or message.get("kind") != "send_message":
+                continue
+            sender = message.get("sender")
+            recipient = message.get("recipient")
+            if not isinstance(sender, str) or not sender.startswith("antigravity:"):
+                continue
+            if not isinstance(recipient, str) or not recipient.startswith("antigravity:"):
+                continue
+            sender_id = sender.removeprefix("antigravity:")
+            child_id = recipient.removeprefix("antigravity:")
+            child_entry = children.get(child_id)
+            if child_entry is None:
+                continue
+            parent_scope, parent_path = topology[child_id]
+            if sender_id != parent_scope:
+                # Exact sibling/other-agent messages are messages, not parent tasks.
+                continue
+            child_preview = child_entry["conversation_preview"]
+            task = dict(message)
+            task.update(
+                {
+                    "kind": "task",
+                    "phase": "assignment",
+                    "sender": parent_path,
+                    "recipient": str(child_preview.get("agent_path") or ""),
+                    "content_role": "antigravity_addressed_task",
+                    "provider_sender_id": sender_id,
+                    "provider_recipient_id": child_id,
+                    "delivery_observed": False,
+                    "consumption_observed": False,
+                }
+            )
+            additions[child_id].append(task)
+
+    for child_id, tasks in additions.items():
+        if not tasks:
+            continue
+        preview = children[child_id]["conversation_preview"]
+        combined = [
+            dict(message)
+            for message in preview.get("messages") or []
+            if isinstance(message, dict)
+        ] + tasks
+        combined.sort(
+            key=lambda message: (
+                str(message.get("timestamp") or ""),
+                message.get("ordinal")
+                if isinstance(message.get("ordinal"), int)
+                else 2**63 - 1,
+            )
+        )
+        seen: set[tuple[object, ...]] = set()
+        messages: list[dict[str, Any]] = []
+        for message in combined:
+            key = _history_message_key(message)
+            if key in seen:
+                continue
+            seen.add(key)
+            messages.append(message)
+        preview["message_count"] = len(messages)
+        preview["messages_truncated"] = False
+        preview["messages"] = messages
+
+
 def _restore_complete_histories(
     entries: list[dict[str, Any]],
     snapshots: list[tuple[dict[str, Any], dict[str, Any]]],
@@ -430,6 +547,7 @@ def conversation_record_entries(
 ) -> list[dict[str, Any]]:
     """Publish root-only Ollama turns under the one run root without changing raw evidence."""
     entries = _core_conversation_record_entries(graph, run_root)
+    _project_antigravity_addressed_tasks(entries, graph)
     root_id = _ollama_root_agent_id(graph)
     if root_id is None:
         return entries

{execweave-0.8.5 → execweave-0.8.6}/src/execweave/viewer_agent_panel.py

@@ -74,6 +74,17 @@ function recordFor(node){
 function canonicalRootRecord(){
   const roots=entries.filter(entryHasRootAuthority);
   const sourceIds=[...new Set(roots.map(entry=>String(entry?.source_id||'')).filter(Boolean))];
+  if(sourceIds.length!==1){
+    if(sourceIds.length)return null;
+    const agy=entries.filter(entry=>{
+      const preview=entry?.conversation_preview;
+      return String(entry?.provider||'').toLowerCase()==='antigravity'&&!!preview&&
+        String(entry?.source_id||'').startsWith('agent:antigravity:conversation:')&&
+        !String(preview.parent_agent_path||'').trim();
+    });
+    const agyIds=[...new Set(agy.map(entry=>String(entry?.source_id||'')).filter(Boolean))];
+    if(agyIds.length===1)return aggregate(agy.filter(entry=>String(entry?.source_id||'')===agyIds[0]));
+  }
   if(sourceIds.length!==1)return null;
   return aggregate(roots.filter(entry=>String(entry?.source_id||'')===sourceIds[0]));
 }

{execweave-0.8.5 → execweave-0.8.6}/src/execweave/viewer_dashboard_focus.py

@@ -39,14 +39,20 @@ execweaveDashboardGraph=function(data){
   });
   const presentationAlias=new Map();
   const antigravityRoot=prepared.find(node=>String(node?.id||'')==='agent:Antigravity');
-  const antigravityMains=prepared.filter(node=>{
+  const antigravityScoped=prepared.filter(node=>{
     const attrs=node?.attributes||{};
     return node?.type==='agent'&&String(attrs.provider||'').toLowerCase()==='antigravity'&&
-      String(node.id||'').startsWith('agent:antigravity:conversation:')&&!String(attrs.parent_agent_path||'').trim();
+      String(node.id||'').startsWith('agent:antigravity:conversation:');
   });
+  const parentScopes=new Set(antigravityScoped.filter(node=>String(node?.attributes?.parent_agent_path||'').trim()).map(node=>String(node?.attributes?.parent_scope_id||'')).filter(Boolean));
+  const evidenceMains=antigravityScoped.filter(node=>{const attrs=node?.attributes||{};return !String(attrs.parent_agent_path||'').trim()&&parentScopes.has(String(attrs.conversation_id||''));});
+  const fallbackMains=antigravityScoped.filter(node=>{const attrs=node?.attributes||{};return !String(attrs.parent_agent_path||'').trim()&&!attrs.routing_identity_only;});
+  const antigravityMains=evidenceMains.length?evidenceMains:fallbackMains;
   if(antigravityRoot&&antigravityMains.length===1){
     const main=antigravityMains[0];presentationAlias.set(antigravityRoot.id,main.id);
     prepared=prepared.filter(node=>node.id!==antigravityRoot.id).map(node=>node.id===main.id?{...node,name:'/root'}:node);
+  }else if(antigravityRoot&&antigravityScoped.length){
+    prepared=prepared.filter(node=>node.id!==antigravityRoot.id);
   }
   const ollamaRoots=prepared.filter(node=>node?.type==='agent'&&['agent:Ollama','agent:ollama'].includes(String(node.id||'')));
   const ollamaRuntimes=prepared.filter(node=>node?.type==='model_runtime'&&String(node?.attributes?.provider||'').toLowerCase()==='ollama');

execweave-0.8.6/tests/test_antigravity_root_child_history_v2.py

@@ -0,0 +1,173 @@
+from __future__ import annotations
+
+import json
+import os
+from pathlib import Path
+
+import pytest
+
+from execweave.antigravity_full_fidelity import antigravity_hook_to_content_events
+from execweave.content_store import FullFidelityContentStore
+from execweave.conversation_records import conversation_record_entries
+from execweave.graph import GraphAccumulator
+from execweave.viewer_projection import write_graph_html
+
+ROOT = "root-current-wire"
+CHILD_A = "child-current-a"
+CHILD_B = "child-current-b"
+ROOT_ID = f"agent:antigravity:conversation:{ROOT}"
+CHILD_A_ID = f"agent:antigravity:conversation:{CHILD_A}"
+CHILD_B_ID = f"agent:antigravity:conversation:{CHILD_B}"
+
+
+def _brain_transcript(tmp_path: Path, conversation_id: str) -> Path:
+    path = tmp_path / "brain" / conversation_id / ".system_generated" / "logs" / "transcript.jsonl"
+    path.parent.mkdir(parents=True, exist_ok=True)
+    return path
+
+
+def _write(path: Path, rows: list[dict[str, object]]) -> None:
+    path.write_text("".join(json.dumps(row, ensure_ascii=False) + "\n" for row in rows), encoding="utf-8")
+
+
+def _root_rows(tmp_path: Path) -> tuple[list[dict[str, object]], list[dict[str, str]]]:
+    specs = [
+        {"Model": "inherit", "Prompt": "child A first task", "Role": "Role A", "TypeName": "worker", "Workspace": "inherit"},
+        {"Model": "inherit", "Prompt": "child B first task", "Role": "Role B", "TypeName": "worker", "Workspace": "inherit"},
+    ]
+    result = (
+        "Created the following subagents:\n"
+        + json.dumps({"conversationId": CHILD_A, "logAbsoluteUri": _brain_transcript(tmp_path, CHILD_A).as_uri(), "workspaceUris": [tmp_path.as_uri()]})
+        + "\n"
+        + json.dumps({"conversationId": CHILD_B, "logAbsoluteUri": _brain_transcript(tmp_path, CHILD_B).as_uri(), "workspaceUris": [tmp_path.as_uri()]})
+    )
+    rows: list[dict[str, object]] = [
+        {"step_index": 0, "source": "USER_EXPLICIT", "type": "USER_INPUT", "status": "DONE", "created_at": "2026-09-01T01:00:00Z", "content": "<USER_REQUEST>root first request</USER_REQUEST>"},
+        {"step_index": 1, "source": "MODEL", "type": "PLANNER_RESPONSE", "status": "DONE", "created_at": "2026-09-01T01:00:01Z", "tool_calls": [{"name": "invoke_subagent", "args": {"Subagents": specs}}]},
+        {"step_index": 2, "source": "MODEL", "type": "GENERIC", "status": "DONE", "created_at": "2026-09-01T01:00:02Z", "content": result},
+        {"step_index": 3, "source": "MODEL", "type": "PLANNER_RESPONSE", "status": "DONE", "created_at": "2026-09-01T01:00:03Z", "content": "root first response"},
+        {"step_index": 4, "source": "USER_EXPLICIT", "type": "USER_INPUT", "status": "DONE", "created_at": "2026-09-01T01:00:50Z", "content": "<USER_REQUEST>root second request</USER_REQUEST>"},
+        {"step_index": 5, "source": "MODEL", "type": "PLANNER_RESPONSE", "status": "DONE", "created_at": "2026-09-01T01:00:51Z", "content": "root dispatches a new child task"},
+    ]
+    return rows, specs
+
+
+def _build(tmp_path: Path) -> tuple[dict[str, object], Path]:
+    run_root = tmp_path / "run"
+    store = FullFidelityContentStore(run_root)
+    root_path = _brain_transcript(tmp_path, ROOT)
+    root_rows, specs = _root_rows(tmp_path)
+    _write(root_path, root_rows)
+    _write(_brain_transcript(tmp_path, CHILD_A), [
+        {"step_index": 0, "source": "MODEL", "type": "PLANNER_RESPONSE", "status": "DONE", "created_at": "2026-09-01T01:00:05Z", "content": "child A first response"},
+        {"step_index": 4, "source": "MODEL", "type": "PLANNER_RESPONSE", "status": "DONE", "created_at": "2026-09-01T01:01:05Z", "content": "child A second response"},
+    ])
+    _write(_brain_transcript(tmp_path, CHILD_B), [
+        {"step_index": 0, "source": "MODEL", "type": "PLANNER_RESPONSE", "status": "DONE", "created_at": "2026-09-01T01:00:06Z", "content": "child B first response"},
+    ])
+    graph = GraphAccumulator(session_id="agy-root-child-v2", source_path=run_root / "events.jsonl")
+    invoke_payload = {"conversationId": ROOT, "workspacePaths": [str(tmp_path)], "transcriptPath": str(root_path), "stepIdx": 1, "toolCall": {"name": "invoke_subagent", "args": {"Subagents": specs}}}
+    for event in antigravity_hook_to_content_events(invoke_payload, hook_event="PostToolUse", store=store, timestamp="2026-09-01T01:00:04Z"):
+        graph.apply(event)
+    root_stop = {"conversationId": ROOT, "workspacePaths": [str(tmp_path)], "transcriptPath": str(root_path), "executionNum": 1, "terminationReason": "done", "fullyIdle": True}
+    for event in antigravity_hook_to_content_events(root_stop, hook_event="Stop", store=store, timestamp="2026-09-01T01:03:00Z"):
+        graph.apply(event)
+    for child_id in (CHILD_A, CHILD_B):
+        stop_payload = {"conversationId": child_id, "workspacePaths": [str(tmp_path)], "transcriptPath": str(_brain_transcript(tmp_path, child_id)), "executionNum": 1, "terminationReason": "done", "fullyIdle": True}
+        for event in antigravity_hook_to_content_events(stop_payload, hook_event="Stop", store=store, timestamp="2026-09-01T01:02:00Z"):
+            graph.apply(event)
+    followup = {"conversationId": ROOT, "stepIdx": 10, "workspacePaths": [str(tmp_path)], "toolCall": {"name": "send_message", "args": {"Recipient": CHILD_A, "Message": "child A second task"}}}
+    for event in antigravity_hook_to_content_events(followup, hook_event="PostToolUse", store=store, timestamp="2026-09-01T01:01:00Z"):
+        graph.apply(event)
+    sibling = {"conversationId": CHILD_B, "stepIdx": 11, "workspacePaths": [str(tmp_path)], "toolCall": {"name": "send_message", "args": {"Recipient": CHILD_A, "Message": "sibling note, not a task"}}}
+    for event in antigravity_hook_to_content_events(sibling, hook_event="PostToolUse", store=store, timestamp="2026-09-01T01:01:01Z"):
+        graph.apply(event)
+    materialized = graph.to_dict()
+    materialized["nodes"].append({"id": "agent:Antigravity", "type": "agent", "name": "Antigravity", "attributes": {"provider": "antigravity"}})
+    return materialized, run_root
+
+
+def _entry(entries: list[dict[str, object]], source_id: str) -> dict[str, object]:
+    matches = [entry for entry in entries if entry.get("source_id") == source_id and isinstance(entry.get("conversation_preview"), dict)]
+    assert len(matches) == 1, (source_id, len(matches))
+    return matches[0]
+
+
+def test_current_generic_result_and_model_field_produce_positive_child_topology(tmp_path: Path) -> None:
+    graph, run_root = _build(tmp_path)
+    for child_id in (CHILD_A_ID, CHILD_B_ID):
+        child = next(node for node in graph["nodes"] if node["id"] == child_id)
+        attrs = child["attributes"]
+        assert attrs["parent_agent_path"] == "/root"
+        assert attrs["parent_scope_id"] == ROOT
+        assert attrs["parent_relation_source"] == "provider_validated_child_transcript"
+    entries = conversation_record_entries(graph, run_root)
+    preview_a = _entry(entries, CHILD_A_ID)["conversation_preview"]
+    texts_a = [message["text"] for message in preview_a["messages"]]
+    assert texts_a == ["child A first task", "child A first response", "child A second task", "child A second response"]
+    second = next(message for message in preview_a["messages"] if message["text"] == "child A second task")
+    assert second["phase"] == "assignment"
+    assert second["content_role"] == "antigravity_addressed_task"
+    assert second["provider_sender_id"] == ROOT
+    assert second["provider_recipient_id"] == CHILD_A
+    assert second["delivery_observed"] is False
+    assert second["consumption_observed"] is False
+    assert "sibling note, not a task" not in texts_a
+    preview_b = _entry(entries, CHILD_B_ID)["conversation_preview"]
+    assert "child A second task" not in [message["text"] for message in preview_b["messages"]]
+
+
+def _required_browser(playwright: object):
+    try:
+        return playwright.chromium.launch()
+    except Exception as error:  # noqa: BLE001
+        if os.environ.get("EXECWEAVE_E2E_REQUIRED", "").lower() not in {"", "0", "false"}:
+            pytest.fail(f"Chromium required for Antigravity root/child history gate: {error}")
+        pytest.skip(f"Chromium unavailable: {error}")
+
+
+@pytest.mark.viewer_e2e
+def test_dashboard_has_one_root_and_reused_child_has_folded_history(tmp_path: Path) -> None:
+    sync_api = pytest.importorskip("playwright.sync_api")
+    graph, run_root = _build(tmp_path)
+    entries = conversation_record_entries(graph, run_root)
+    viewer = run_root / "viewer.html"
+    write_graph_html(graph, viewer)
+    with sync_api.sync_playwright() as playwright:
+        browser = _required_browser(playwright)
+        try:
+            page = browser.new_page(viewport={"width": 1440, "height": 1000})
+            page.goto(viewer.as_uri())
+            page.wait_for_selector(".node")
+            visible = page.eval_on_selector_all(".node", "nodes=>nodes.map(node=>({id:node.dataset.id,text:node.textContent}))")
+            ids = [item["id"] for item in visible]
+            assert "agent:Antigravity" not in ids
+            assert ROOT_ID in ids
+            assert "/root" in next(item for item in visible if item["id"] == ROOT_ID)["text"]
+            assert "Role A" in next(item for item in visible if item["id"] == CHILD_A_ID)["text"]
+            assert "Role B" in next(item for item in visible if item["id"] == CHILD_B_ID)["text"]
+            page.eval_on_selector(f'.node[data-id="{CHILD_A_ID}"]', "node=>node.dispatchEvent(new MouseEvent('click',{bubbles:true}))")
+            page.wait_for_function("()=>(document.getElementById('details')?.innerText||'').includes('child A second task')")
+            details = page.locator("#details")
+            current = details.inner_text()
+            assert "child A second task" in current
+            assert "child A second response" in current
+            assert "sibling note, not a task" not in current
+            assert details.locator(".execweave-agent-older").count() == 1
+            older = details.locator(".execweave-agent-older").first
+            assert not older.evaluate("node=>node.open")
+            older.locator("summary").click()
+            assert older.evaluate("node=>node.open")
+            expanded = older.inner_text()
+            assert "child A first task" in expanded
+            assert "child A first response" in expanded
+            assert "child A second task" not in expanded
+            assert "child A second response" not in expanded
+            page.evaluate("items=>window.__execweaveAgentPanel.setEntries(items)", entries)
+            persisted = details.locator(".execweave-agent-older").first
+            assert persisted.evaluate("node=>node.open")
+            persisted_text = persisted.inner_text()
+            assert "child A first task" in persisted_text
+            assert "child A first response" in persisted_text
+        finally:
+            browser.close()

{execweave-0.8.5 → execweave-0.8.6}/tests/test_v069_dashboard_release.py

@@ -103,10 +103,10 @@ def test_release_version_and_noncommercial_license_metadata_are_078() -> None:
     stage-integrity node-ID floor refuses renames; the assertions below are the
     current release.
     """
-    assert __version__ == "0.8.5"
+    assert __version__ == "0.8.6"
     pyproject = Path("pyproject.toml").read_text(encoding="utf-8")
-    assert 'version = "0.8.5"' in pyproject
-    assert "ExecWeave v0.8.5 release metadata" in pyproject
+    assert 'version = "0.8.6"' in pyproject
+    assert "ExecWeave v0.8.6 release metadata" in pyproject
     assert "License :: Other/Proprietary License" in pyproject
     assert "License :: OSI Approved :: MIT License" not in pyproject