MCP: remove superseded Python facade and mesh

backend/app/__main__.py

@@ -105,22 +105,6 @@ def main():
     except Exception as exc:
         log.warning("⚠️  Startup health check error: %s", exc)
 
-    # MCP Integrity Check (fast structural only)
-    try:
-        from .api.mcp_guardrails import validate_mcp_integrity
-        report = validate_mcp_integrity()
-        if not report["ok"]:
-            log.warning("⚠️  MCP integrity check FAILED: %s", report["errors"])
-            log.warning("   Run 'skill_mcp_self_heal' to attempt auto-repair")
-        else:
-            log.info("✅ MCP integrity check passed")
-    except Exception as exc:
-        log.warning("⚠️  MCP integrity check error: %s", exc)
-
-    # ── Auto-Start Hivemind (port 8490) ────────────────────────
-    from .mcp.hivemind_launcher import start_hivemind
-    start_hivemind()
-
     # Delegate to snackbar
     from .core.snackbar import main as run_snackbar
     run_snackbar()

backend/app/api/chat.py

@@ -321,7 +321,7 @@ async def _execute_tool(tool_name: str, arguments: dict) -> str:
         return json.dumps({
             "ollama": {"status": "checking"},
             "backend": {"status": "online"},
-            "mcp_servers": {"count": "check /api/mcp/diagnostics for details"},
+            "mcp_gateway": {"mode": "external stdio client"},
         })
 
     if tool_name == "binder_create":

backend/app/api/developer_api.py

@@ -1031,11 +1031,9 @@ async def handle_developer_chat(request: web.Request) -> web.Response:
             "• /api/developer/repos/{name}/file-preview?path=... — preview file content\n"
             "• /api/developer/repos/{name}/stage — stage a file (POST)\n"
             "• /api/developer/repos/{name}/commit — commit staged files (POST)\n\n"
-            "**Skills & MCP:**\n"
-            "• /api/skills — list all 54+ built-in skills\n"
-            "• /api/skills/{skill_id}/run — execute a named skill\n"
-            "• /api/mcp/tools — list MCP server tools\n"
-            "• /api/mcp/diagnostics — MCP health diagnostics\n\n"
+            "**Skills:**\n"
+            "• /api/skills — list governed built-in skills\n"
+            "• /api/skills/{skill_id}/run — execute a named skill\n\n"
             "**Health & System:**\n"
             "• /api/control/status — full ecosystem health (Ollama, Hivemind, providers, etc.)\n"
             "• /api/ollama/status — Ollama model status\n"