feat: deep-agent canvas, live observability, and multi-environment tooling

Self-hosted platform for building, testing, and shipping LangChain/LangGraph agents. Deep-agent sub-agents on the canvas, a live tracing/observability timeline, auto-provisioned built-in tools with import/export, per-environment tool variables, streamed evaluations, and per-user auth token forwarding.

apps/web/components/markdown.tsx

@@ -0,0 +1,26 @@
+"use client";
+/* Renders an assistant reply as GitHub-Flavored Markdown (Feature 1 - structured responses).
+   Safe by default: react-markdown does NOT render raw HTML (no rehype-raw), so agent output
+   cannot inject markup. Visual styling lives in the `.md` block in app/globals.css and uses
+   the app's design tokens, so it adapts to light/dark automatically. Memoized so finalized
+   messages don't re-parse while a newer message is still streaming. */
+import { memo } from "react";
+import ReactMarkdown from "react-markdown";
+import remarkGfm from "remark-gfm";
+
+export const Markdown = memo(function Markdown({ children }: { children: string }) {
+  return (
+    <div className="md">
+      <ReactMarkdown
+        remarkPlugins={[remarkGfm]}
+        components={{
+          a: ({ node, ...props }: any) => (
+            <a {...props} target="_blank" rel="noopener noreferrer" />
+          ),
+        }}
+      >
+        {children}
+      </ReactMarkdown>
+    </div>
+  );
+});