feat: add persistent Deno cache to Railway executor for faster tool loading

- Set DENO_DIR=/app/.deno_cache to persist module cache
- Pre-cache common dependencies (zod-to-json-schema, ai, zod) during build
- Add Railway volume configuration for /app/.deno_cache
- Improve logging to show cache hits vs network downloads
- Add --allow-read and --allow-write permissions for cache access

This dramatically reduces tool loading time after the first import.
Dependencies are downloaded once and reused across all subsequent requests.

Example: ctx-zip with 200+ dependencies will only download once instead
of on every chat request.

🤖 Generated with [Claude Code](https://claude.com/claude-code)

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

apps/railway-executor/Dockerfile

@@ -4,11 +4,27 @@ FROM denoland/deno:1.39.0
 # Set working directory
 WORKDIR /app
 
-# Copy the server file
+# Set Deno cache directory to persist across requests
+ENV DENO_DIR=/app/.deno_cache
+
+# Create cache directory
+RUN mkdir -p /app/.deno_cache
+
+# Copy files
 COPY server.ts .
+COPY deno.json .
+
+# Pre-cache zod-to-json-schema (used by server.ts)
+RUN deno cache --reload \
+  https://esm.sh/zod-to-json-schema@3.25.0
+
+# Pre-cache common AI SDK dependencies to speed up first loads
+RUN deno cache --reload \
+  https://esm.sh/ai@5.0.83 \
+  https://esm.sh/zod@3.22.0 || true
 
 # Expose port (Railway will set PORT env var)
 EXPOSE 3002
 
 # Run the Deno server
-CMD ["deno", "run", "--allow-net", "--allow-env", "server.ts"]
+CMD ["deno", "run", "--allow-net", "--allow-env", "--allow-read", "--allow-write", "server.ts"]