refactor(viz): use AP typed tools (get_context/get_impact) + bump 3.18.1

Per architect plan — refactor the impact path to fully use automatised-pipeline's
typed MCP tools instead of hand-written Cypher (no new module; behavior-preserving):
- _impact_for_graph: symbol-level direction now from get_context per member
  (cap N=20, Cypher fallback beyond) — gains implements/implemented_by +
  uses/used_by + community, on top of calls/imports. Blast-radius counts
  (communities_affected/processes_affected) from get_impact. File→File edges,
  present-gate, members list, _rollup kept as legitimate query_graph use.
- _ast_and_impact: per-symbol get_context (capped) + one get_impact.
- workflow_graph_source_ast: search_codebase raw call → typed wrapper.
All existing response keys preserved; only additive fields. Frontend reads no
ast.impact array (verified) — safe.

Version 3.18.0 → 3.18.1 across ALL marketplace files (pyproject, plugin.json,
marketplace.json metadata+plugin+version_note, README badge) — shipping the
since-3.18.0 fixes: HUD true counts (/api/stats), AP-typed-tools impact,
ap_bridge installed_plugins.json resolution, HTML-attr XSS + node-id hash.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

Author: cdeust

.claude-plugin/marketplace.json

@@ -6,22 +6,22 @@
   },
   "metadata": {
     "description": "Persistent memory and cognitive profiling plugins for Claude Code",
-    "version": "3.18.0"
+    "version": "3.18.1"
   },
   "plugins": [
     {
       "name": "cortex",
       "source": "./",
       "description": "Persistent memory and cognitive profiling for Claude Code — thermodynamic memory with heat/decay, intent-aware retrieval, biological plasticity, codebase intelligence, and cognitive profiling. 49 MCP tools with enriched schemas. PostgreSQL + pgvector in CLI mode; automatic SQLite fallback in Cowork/sandboxed mode. v3.17.0 — autonomous per-project wiki: SessionStart auto-spawns a 6-hour consolidate cycle; a headless `claude -p` worker drains the curation-gap queue, calls codebase-intelligence MCP tools to ground each section in the real call graph, and authors missing anchor pages (architecture / services / api / data-flow / operations / decisions / PRD) per project from the source tree. 15 canonical scopes × 13 file sections; per-project dashboards under `wiki/_dashboards/`. Mermaid diagrams have a 🔍 lens with zoom + pan. Workflow graph with caller-qualified CALLS chains rendering full method-to-method dependencies (native tree-sitter, no AP required). Side panel humanized for non-technical users. Ingests codebase analysis (ai-automatised-pipeline) and PRDs (prd-spec-generator) into wiki + memory + knowledge graph. Docker image available.",
-      "version": "3.18.0",
+      "version": "3.18.1",
       "author": {
         "name": "Clement Deust",
         "email": "admin@ai-architect.tools"
       },
       "homepage": "https://github.com/cdeust/Cortex",
       "repository": "https://github.com/cdeust/Cortex",
       "license": "MIT",
-      "version_note": "3.18.0 — live execution-trace graph in cortex-visualize: domain→session→chain served live from session transcripts (no snapshot), galaxy-style non-overlapping session clusters with type-banded disks, and a file impact panel wired to the automatised-pipeline code graph for full caller/callee direction, causal execution chains (processes), and per-file git version history. Loopback-only host guard retained (GHSA-gvpp-v77h-5w8g).",
+      "version_note": "3.18.1 — HUD shows true store counts (memories/entities/synapses) via /api/stats instead of loaded-graph nodes; impact panel refactored to use automatised-pipeline's typed tools (get_context/get_impact) for full direction incl. implements/uses + community/process blast radius; AP MCP resolved from installed_plugins.json (compiled binary, not bin/* glob); HTML-attribute XSS escaping + non-security node-id hash. 3.18.0 — live execution-trace graph in cortex-visualize: domain→session→chain (no snapshot), galaxy-style type-banded clusters, file impact panel wired to the AP code graph (caller/callee direction, causal chains, per-file git versions). Loopback-only host guard retained (GHSA-gvpp-v77h-5w8g).",
       "keywords": ["memory", "cognitive-profiling", "mcp", "claude-code", "cortex", "knowledge-graph", "codebase-analysis", "docker"],
       "category": "productivity",
       "runtime": ["cli", "cowork"],

.claude-plugin/plugin.json

@@ -1,7 +1,7 @@
 {
   "name": "cortex",
   "description": "Persistent memory for Claude Code — remembers across sessions automatically. Install and forget. Scientific retrieval backed by 41 published papers.",
-  "version": "3.18.0",
+  "version": "3.18.1",
   "author": {
     "name": "Clement Deust",
     "email": "admin@ai-architect.tools"

README.md

@@ -8,7 +8,7 @@
   <img src="https://img.shields.io/badge/python-3.10+-blue.svg" alt="Python 3.10+">
   <img src="https://img.shields.io/badge/tests-2500_passing-brightgreen.svg" alt="Tests">
   <img src="https://img.shields.io/badge/citations-45_papers-orange.svg" alt="Citations">
-  <img src="https://img.shields.io/badge/version-3.18.0-brightgreen.svg" alt="Version 3.18.0">
+  <img src="https://img.shields.io/badge/version-3.18.1-brightgreen.svg" alt="Version 3.18.1">
   <a href="https://glama.ai/mcp/servers/cdeust/Cortex"><img src="https://glama.ai/mcp/servers/cdeust/Cortex/badges/score.svg" alt="Glama score: security A, license A"></a>
 </p>
 

mcp_server/infrastructure/workflow_graph_source_ast.py

@@ -425,10 +425,7 @@ def search_codebase(
         if not gp:
             return []
         resp = self._loop_owner.run(
-            self._bridge.call(
-                "search_codebase",
-                {"graph_path": gp, "query": query, "limit": int(limit)},
-            )
+            self._bridge.search_codebase(gp, query, limit=int(limit))
         )
         out: list[dict[str, Any]] = []
         for r in _as_list(resp):