Whoami User Identity Profile Management

The Problem

AI agents usually start each new session without knowing who the user is: tech stack, preferences, and common tools must be re-explained, and different AI tools do not share user context. whoami stores a Markdown user profile remotely so an agent can read identity information during the first interaction and then continue with the user’s actual task.

How It Works

The skill centers on reading, updating, and sharing a remote profile:
- get: fetches the remote profile; outputs Markdown when one exists, or prompts creation when missing.
- update: overwrites the remote profile with complete content, preferably via a temporary file and the --file flag.
- metadata: reports profile metadata such as character count and line count.
- Cross-tool sharing: multiple AI tools can read the same profile, reducing repeated self-introductions.

On first use, the script opens a browser login page. The user generates a one-time configuration URL from the Dashboard, and the agent saves the URL content to ~/.whoamiagent. After loading the profile, the agent should use it as context and proceed with coding, calling other skills, or other requested work instead of stopping at identity loading.

Boundaries

• update is an overwrite operation and requires the full profile; the remote side retains the last 3 versions.
• Profile content must stay under 5000 characters, so it should be concise.
• Setup depends on a remote service and browser login, and the agent should receive a one-time URL rather than handling API keys directly.
• It fits the “who is the user and what do they prefer” context problem, not large files, complex org structures, or real-time permission systems.