PilotMyX

What is the fastest setup path?

Open MCP Access, create a scoped token, download the PilotMyX skills bundle, then copy the universal setup prompt into your agent.

The setup prompt contains the public Streamable HTTP endpoint and the bundle download URL. Your token is shown once and must be stored through the agent's secure environment or secret mechanism, never in a tracked configuration file.

Let the agent use its native MCP configuration flow. Once configured, ask it to list the available PilotMyX tools and inspect account freshness before doing any analysis.

• Give the token a recognizable name
• Select only the permissions the agent needs
• Choose an expiry for temporary access
• Copy the secret before closing the token dialog
• Install the required MCP skill and the optional Voice skill if wanted
• Paste the universal setup prompt
• Verify the connection with a read-only request first

What do the two PilotMyX skills add?

The bundle separates the required MCP operating instructions from optional local voice memory.

The PilotMyX MCP skill is self-contained. It teaches the agent how to inspect freshness, query representative samples, compare posts and replies, plan from evidence, and use draft or publication tools safely.

The PilotMyX Voice skill is optional and independent. It learns from published examples and explicit feedback, then maintains voice rules, anchors, lessons, and experiments in a local notes file. PilotMyX does not store this memory or generate content behind the scenes.

Both skills are included in the current download, but they install as separate directories. This lets the same PilotMyX account work with Codex, Claude, or another compatible client without tying strategy to a proprietary server-side model.

Which permissions should I enable?

Start read-only, then add draft or publishing permissions only when the workflow requires them.

Give each agent only the access it needs. An analytics workflow can stay read-only, while a publishing workflow can add drafts, scheduling, or immediate publishing.

• Read account data and analytics
• Read drafts and scheduled publications
• Create or edit drafts and media
• Schedule, reschedule, or cancel future publications
• Publish immediately

How do I keep MCP access safe?

Treat every MCP token like a password and keep write permissions narrow.

• Use a separate token for each agent or device
• Never place a token in a screenshot, post, prompt history, or tracked file
• Ask the agent to confirm before every write or publish action
• Revoke and replace a token immediately if it may have been exposed

Can I change or remove access?

Revoke a token from MCP Access at any time, then create a replacement with the permissions you want.

Tokens are independent. Replacing one agent's credential does not disconnect the others, and revocation does not require changing your X password.