v0.0.7

• Add Python-backed skills that install into the persistent IPython kernel and appear alongside markdown skills.
• Add Prime team selection during Prime Inference login so team inference costs use the selected Prime CLI context.
• Guide unauthenticated users through login and model selection before the first agent turn.
• Use a bounded animated Prime Lab splash with centered progress and hidden command output in the installer.
• Fix Prime CLI credentials from prime login to make Prime Inference models available on startup, and quiet stale notices during onboarding.
• Dependency bumps.