release: v2.6.0 — remove memory dashboard, global memory dedup fix, cleanup

- Remove 3D memory dashboard (brain graph) entirely — only unified 2D
  neural graph remains
- Remove open_memory_dashboard tool (34 → 33 tools)
- Delete unused 3D leftover files from unified viz (clusters, layout,
  scene, nodes, edges, effects, interaction, batch, zoom)
- Remove domain clustering force from unified graph renderer
- Fix global memory edge explosion: deduplicate domain_hub_ids by hub_id
  so global nodes connect to ~10 unique hubs instead of 61 aliases
- Update all version references to 2.6.0
- Clean up docs, skills, README, CLAUDE.md references
- ruff format all files

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

Author: cdeust

.claude-plugin/marketplace.json

@@ -6,14 +6,14 @@
   },
   "metadata": {
     "description": "Persistent memory and cognitive profiling plugins for Claude Code",
-    "version": "2.5.5"
+    "version": "2.6.0"
   },
   "plugins": [
     {
       "name": "cortex",
       "source": "./",
       "description": "Persistent memory and cognitive profiling for Claude Code — thermodynamic memory, intent-aware retrieval, codebase intelligence, knowledge graphs, and cognitive profiling. Docker image available.",
-      "version": "2.5.5",
+      "version": "2.6.0",
       "author": {
         "name": "Clement Deust",
         "email": "admin@ai-architect.tools"

.claude-plugin/plugin.json

@@ -1,7 +1,7 @@
 {
   "name": "cortex",
   "description": "Persistent memory and cognitive profiling for Claude Code — thermodynamic memory with heat/decay, intent-aware retrieval, biological plasticity, codebase intelligence, and cognitive profiling. 35 MCP tools. Docker image available with PostgreSQL + pgvector pre-configured.",
-  "version": "2.5.5",
+  "version": "2.6.0",
   "author": {
     "name": "Clement Deust",
     "email": "admin@ai-architect.tools"

CLAUDE.md

@@ -210,16 +210,15 @@ Handlers are the **composition roots**: they wire infrastructure (I/O) to core (
 - `embedding_engine.py` — Vector embeddings (384-dim, sentence-transformers)
 - `agent_config.py` — Agent configuration and topic scoping
 
-**handlers/** — Composition roots (34 tools + helpers, one per tool)
+**handlers/** — Composition roots (33 tools + helpers, one per tool)
 
 **validation/** — `schemas.py` — Per-tool argument validation
 
 **errors/** — `__init__.py` — MethodologyError, ValidationError, StorageError, AnalysisError, McpConnectionError
 
-**server/** — HTTP servers and visualization (5 modules)
+**server/** — HTTP servers and visualization (4 modules)
 - `http_server.py` — Visualization HTTP server
-- `http_viz_server.py` — 3D neural graph visualization server
-- `http_dashboard_server.py` — Memory dashboard server
+- `http_viz_server.py` — Unified neural graph visualization server
 - `http_dashboard_data.py` — Dashboard data aggregation
 - `http_common.py` — Shared HTTP utilities
 
@@ -231,7 +230,7 @@ Handlers are the **composition roots**: they wire infrastructure (I/O) to core (
 
 ## MCP Tools
 
-### Tier 1 — Core Memory & Profiling (22 tools)
+### Tier 1 — Core Memory & Profiling (21 tools)
 
 | Tool | Purpose | Target Latency |
 |---|---|---|
@@ -243,7 +242,6 @@ Handlers are the **composition roots**: they wire infrastructure (I/O) to core (
 | `get_methodology_graph` | Graph data for 3D visualization | <100ms |
 | `open_visualization` | Launch 3D methodology map in browser | — |
 | `explore_features` | Interpretability exploration (features, attribution, persona, crosscoder) | <100ms |
-| `open_memory_dashboard` | Launch real-time memory dashboard in browser | — |
 | `remember` | Store a memory through the 4-signal predictive coding gate | <100ms |
 | `recall` | Retrieve memories via 6-signal WRRF fusion | <200ms |
 | `consolidate` | Run maintenance: decay, compression, CLS, sleep compute | <5s |

README.md

@@ -152,7 +152,7 @@ claude mcp add cortex -- python -m mcp_server
 | `/cortex-profile` | View cognitive style, patterns, blind spots | *"Show my work patterns"* |
 | `/cortex-navigate-knowledge` | Trace relationships, explore causal chains | *"How does auth relate to billing?"* |
 | `/cortex-setup-project` | Bootstrap memory from code or import history | *"Set up Cortex for this project"* |
-| `/cortex-visualize` | Launch 3D neural graph or memory dashboard | *"Show me the memory map"* |
+| `/cortex-visualize` | Launch unified neural graph visualization | *"Show me the memory map"* |
 | `/cortex-automate` | Create triggers, rules, sync to CLAUDE.md | *"Remind me about X when I open that file"* |
 | `/cortex-debug-memory` | Fix bad memories, rate quality, restore checkpoints | *"That memory is wrong, delete it"* |
 
@@ -271,7 +271,6 @@ Cortex builds a profile of how you work — from your Claude Code session histor
 | `memory_stats` | Memory system diagnostics |
 | `narrative` | Generate project story from stored memories |
 | `import_sessions` | Import conversation history into memory |
-| `open_memory_dashboard` | Launch real-time memory heatmap dashboard |
 
 ### Navigation (5 tools)
 
@@ -356,8 +355,7 @@ Each agent uses `agent_topic` scoping — a **soft boost** that promotes topic-r
 ```bash
 # From any Claude Code session:
 /cortex-visualize    # or directly:
-# cortex:open_visualization    — 3D neural graph at localhost:3458
-# cortex:open_memory_dashboard — Memory heatmap at localhost:3457
+# cortex:open_visualization    — Unified neural graph at localhost:3458
 ```
 
 The graph organizes everything into a 6-level hierarchy — from broad categories down to individual memories and entities. Node size reflects importance, glow reflects heat (recency), and colored arcs show quality scores. Auto-shuts down after 10 minutes idle.

docs/api-reference.md

@@ -140,12 +140,6 @@ Interpretability exploration across four modes.
 
 ---
 
-### `open_memory_dashboard`
-
-Launches the real-time memory dashboard in the default browser. Shows heat map, entity graph, and activity feed.
-
-**Input Schema:** *(no parameters)*
-
 ---
 
 ### `remember`

docs/architecture.md

@@ -193,7 +193,6 @@ Handlers sit at the center as **composition roots**: they wire infrastructure (I
 | `get_methodology_graph.py` | Graph data for 3D visualization |
 | `open_visualization.py` | Launches 3D visualization in browser |
 | `explore_features.py` | 4-mode interpretability exploration |
-| `open_memory_dashboard.py` | Launches real-time memory dashboard |
 
 #### Core Memory
 
@@ -240,7 +239,7 @@ Handlers sit at the center as **composition roots**: they wire infrastructure (I
 | Module | Purpose |
 |---|---|
 | `mcp_router.py` | MCP JSON-RPC dispatch with version negotiation and error boundary |
-| `http_server.py` | Singleton HTTP server for visualization + memory dashboard with idle timeout |
+| `http_server.py` | Singleton HTTP server for visualization with idle timeout |
 
 ### `transport/` — I/O framing
 

docs/safeskill-report.json

@@ -1,6 +1,6 @@
 {
   "packageName": "neuro-cortex-memory",
-  "packageVersion": "2.5.5",
+  "packageVersion": "2.6.0",
   "scanId": "bWrS83FjaYuBcPoLylGid",
   "timestamp": 1774787345549,
   "duration": 1881,

mcp_server/__main__.py

@@ -14,15 +14,16 @@
 
 from fastmcp import FastMCP
 
+from mcp_server import (
+    tool_registry_advanced,
+    tool_registry_core,
+    tool_registry_manage,
+    tool_registry_memory,
+    tool_registry_nav,
+)
 from mcp_server.infrastructure.mcp_client_pool import close_all
 from mcp_server.server.http_server import shutdown_server
-from mcp_server.server.http_dashboard_server import shutdown_memory_dashboard_server
 from mcp_server.server.http_viz_server import shutdown_unified_viz_server
-from mcp_server import tool_registry_core
-from mcp_server import tool_registry_memory
-from mcp_server import tool_registry_manage
-from mcp_server import tool_registry_nav
-from mcp_server import tool_registry_advanced
 
 # ── Server Instance ────────────────────────────────────────────────────────
 
@@ -51,7 +52,6 @@
 def _shutdown(sig=None, frame=None) -> None:
     close_all()
     shutdown_server()
-    shutdown_memory_dashboard_server()
     shutdown_unified_viz_server()
     sys.exit(0)
 

mcp_server/core/ast_extractors_extra.py

@@ -7,8 +7,8 @@
 
 from typing import TYPE_CHECKING
 
+from mcp_server.core.ast_extractors import _find_children, _text, _walk_type
 from mcp_server.core.codebase_parser import ImportInfo, SymbolDef
-from mcp_server.core.ast_extractors import _text, _walk_type, _find_children
 
 if TYPE_CHECKING:
     from tree_sitter import Node

mcp_server/core/ast_parser.py

@@ -128,9 +128,9 @@ def _extract_python(
 ) -> tuple[list[ImportInfo], list[SymbolDef], list[str]]:
     """Extract Python imports, definitions, and call sites."""
     from mcp_server.core.ast_extractors import (
-        extract_python_imports,
-        extract_python_definitions,
         extract_calls_generic,
+        extract_python_definitions,
+        extract_python_imports,
     )
 
     imports = extract_python_imports(root, source)
@@ -148,9 +148,9 @@ def _extract_js(
 ) -> tuple[list[ImportInfo], list[SymbolDef], list[str]]:
     """Extract JavaScript/TypeScript imports, definitions, and calls."""
     from mcp_server.core.ast_extractors import (
-        extract_js_imports,
-        extract_js_definitions,
         extract_calls_generic,
+        extract_js_definitions,
+        extract_js_imports,
     )
 
     imports = extract_js_imports(root, source)
@@ -169,8 +169,8 @@ def _extract_go(
     """Extract Go imports, definitions, and calls."""
     from mcp_server.core.ast_extractors import extract_calls_generic
     from mcp_server.core.ast_extractors_extra import (
-        extract_go_imports,
         extract_go_definitions,
+        extract_go_imports,
     )
 
     return (
@@ -187,8 +187,8 @@ def _extract_swift(
     """Extract Swift imports, definitions, and calls."""
     from mcp_server.core.ast_extractors import extract_calls_generic
     from mcp_server.core.ast_extractors_extra import (
-        extract_swift_imports,
         extract_swift_definitions,
+        extract_swift_imports,
     )
 
     return (
@@ -205,8 +205,8 @@ def _extract_rust(
     """Extract Rust imports, definitions, and calls."""
     from mcp_server.core.ast_extractors import extract_calls_generic
     from mcp_server.core.ast_extractors_extra import (
-        extract_rust_imports,
         extract_rust_definitions,
+        extract_rust_imports,
     )
 
     return (