diff --git a/.agent/workflows/deploy_to_production.md b/.agent/workflows/deploy_to_production.md
index c3a72be..172357f 100644
--- a/.agent/workflows/deploy_to_production.md
+++ b/.agent/workflows/deploy_to_production.md
@@ -3,29 +3,25 @@
 ---
 
 This workflow automates the deployment of the Cortex Hub to the production server located at `192.168.68.113`. 
-The production server uses `cortex-hub` inside `/home/coder/project/cortex-hub`.
+The production server runs the application out of `/home/coder/project/cortex-hub`.
+
+**MAIN KNOWLEDGE POINT:** Agents and Users should refer to `docs/deployment_reference.md` to understand the full proxy and architecture layout prior to running production debugging.
 
 Follow these steps carefully:
 
 1. Sync the local codebase to a temporary directory on the production server via `rsync`.
    * The destination is `/tmp/cortex-hub/`.
-   * We exclude `.git`, `node_modules`, `ui/client-app/node_modules`, and Python cache directories.
-// turbo
+   * Ensure you exclude `.git`, `node_modules`, `__pycache__`, and audio blobs before migrating.
 2. Copy the synced files into the actual project directory replacing the old ones.
-   * We first clean `nginx.conf` and other related files if needed to ensure pure overwrite.
-// turbo
 3. Run the local deployment script on the production server.
    * `cd /home/coder/project/cortex-hub`
    * `sudo bash local_deployment.sh`
 
 ### Automated Script Example
 
-You can also use the single bash script stored in the repository, `remote_deploy.sh`, which performs all these steps.
-
-1. Ensure `sshpass` is installed (`sudo apt-get install sshpass`)
-2. Run `./remote_deploy.sh`
+You should use the single bash script stored in the repository, `remote_deploy.sh`, which safely performs all these steps using secure shell arguments.
 
 ```bash
-# To run this workflow manually via the script:
-bash /app/remote_deploy.sh
+# To safely deploy to production:
+REMOTE_PASS='a6163484a' bash /app/remote_deploy.sh
 ```
diff --git a/docs/deployment_reference.md b/docs/deployment_reference.md
index af59fe4..05203d1 100644
--- a/docs/deployment_reference.md
+++ b/docs/deployment_reference.md
@@ -55,14 +55,22 @@
       3. Performs automated database migrations running parallel idempotent logic (`app/db/migrate.py`) via the `Uvicorn` startup lifecycle.
 
 ### How to Release
-You just run:
+You must safely pass the authentication key into the script from the command line:
 ```bash
-bash /app/remote_deploy.sh
+REMOTE_PASS='a6163484a' bash /app/remote_deploy.sh
 ```
 
 ---
 
-## 3. Reference Troubleshooting & Known Pitfalls
+## 3. Automation Logic for Agents (.agent Workflows)
+For any automated AI attempting to debug or push changes:
+* **Primary Source of Truth**: This `docs/deployment_reference.md` file defines the architecture rules.
+* **Workflows available:** Look at `.agent/workflows/deploy_to_production.md` for explicit directory movement maps.
+* **Envoy Control Map:** Refer to `.agent/workflows/envoy_control_plane.md` if the backend successfully mounts on the remote machine but `ai.jerxie.com/api/` throws 404s/503s.
+
+---
+
+## 4. Reference Troubleshooting & Known Pitfalls
 
 If production encounters a bug or routing fails, these are the historically primary offenders:
 
diff --git a/local_deployment.sh b/local_deployment.sh
index c84def1..f34c634 100644
--- a/local_deployment.sh
+++ b/local_deployment.sh
@@ -56,6 +56,6 @@
 echo "🧹 Cleaning up unused Docker resources..."
 
 # Remove dangling images (images without a tag).
-sudo docker system prune -f || true
+sudo docker image prune -f || true
 
 echo "✨ Cleanup finished."
\ No newline at end of file