Secure Muse Code Gmail Connection with PortEden

Prerequisites

• Muse Code (beta) installed and authenticated: curl -fsSL https://dev.meta.ai/install.sh | bash, then muse login. Muse Code runs on macOS and Linux; on Windows use WSL2.
• A Gmail account you want Muse Code to use. You will authorize it during the flow if your PortEden account does not already have Gmail connected.

Step 1: Add the PortEden MCP server to settings.json

Muse Code configures MCP servers in its user settings file. PortEden runs a hosted MCP server you connect with one URL over the streamable_http transport. Nothing to install, nothing to host.

mkdir -p ~/.config/muse 
$EDITOR ~/.config/muse/settings.json 

{ 
  "schema_version": 1, 
  "mcp_servers": { 
    "gmail": { 
      "transport": "streamable_http", 
      "url": "https://mcp.porteden.com/email", 
      "mode": "optional" 
    } 
  } 
} 

Want more than email later?

Each PortEden capability has its own MCP URL. Add them as additional entries under mcp_servers the same way, each with its own scoped token:

{ 
  "gmail":         "https://mcp.porteden.com/email", 
  "calendar":      "https://mcp.porteden.com/calendar", 
  "drive":         "https://mcp.porteden.com/drive", 
  "google-docs":   "https://mcp.porteden.com/google-docs", 
  "google-sheets": "https://mcp.porteden.com/google-sheets", 
  "tasks":         "https://mcp.porteden.com/tasks" 
} 

Step 2: Sign in to PortEden

The first time Muse Code calls an email tool, the PortEden server requires authentication. This is the single setup screen you will see. There is nothing to do in advance.

Step 3: Connect Gmail (first-time only)

This step runs only if your PortEden account does not already have Gmail connected. If it does, PortEden reuses the existing Gmail connection and you can jump straight to Step 4.

Step 4: Verify the connection

In a Muse Code session, run a low-risk read prompt. Then check the PortEden audit log to confirm the request shows up.

Try one of these

What to confirm

• Muse Code returns real data from your Gmail, not a refusal or an error.
• Sensitive identifiers (full email addresses, phone numbers) appear redacted or tokenized if you left redaction enabled.
• The PortEden audit log at my.porteden.com shows the request with a green allow decision.

Step 5: Tighten what Muse Code can do (optional)

The token PortEden created in Step 2 already uses conservative defaults: email scope only, redaction on, and confirm-before-write for send and delete. Once the connection works end-to-end, tighten or relax it from my.porteden.com under Access Tokens. Remember that every sub-agent Muse Code spawns uses this same token, so its scope is the boundary for the whole fleet.

Permission presets for a coding agent

Recommended rules for a Muse Code token

• Start read-only: Coding workflows rarely need to send. Grant write actions only when a workflow proves it needs them.
• Redaction: Leave on. What the model never receives cannot be retained or trained on, whatever Meta tier the session runs on.
• Contact blocklist: Add HR, legal, and personal aliases (or whole domains) you do not want a coding agent touching.
• Time window: Restrict to messages from the last 90 days unless a workflow needs history.
• Label exclusions: Block Confidential, Legal, and any client-specific labels under NDA.
• Confirm before write: Keep on for send and delete so a preview surfaces before any outbound action.

Suggested prompts for everyday use

Once the connection is verified, these prompts are good starting points for a coding session. Each maps to a small number of PortEden tool calls, so behavior is predictable and the audit log stays clean.

Troubleshooting and error handling

PortEden returns structured errors that Muse Code surfaces in its replies, and Muse Code adds its own settings validation at startup. Match the message you see to the entries below, then jump to the matching debug prompt in the next section.

Debug prompts for Muse Code

When something is wrong but the error message is vague, paste one of these prompts into Muse Code. They make the agent self-report the structured response from PortEden so you can pinpoint the cause without leaving the terminal.

Security best practices

FAQ

Next steps