Personal Squad — Skill Document

Automating AI agent setup across multiple project sessions is achieved by discovering personal agents from a global directory and merging them into active sessions, ensuring consistent advising behavior without manual reconfiguration.

Ambient discovery for personal AI agents is a mechanism that automatically loads agents from a global personal-squad directory when a squad session starts, eliminating the need to manually re-add them.

Personal AI agents cannot modify project state during session merging due to a ghost-protocol read-only constraint, which allows them to advise based on project state while preventing any writes to project-managed state.

You can disable automatic agent discovery for a specific session by setting an environment-variable kill switch, which turns off the ambient discovery process before the session cast is initialized.

Maintaining consistent AI advising behavior across different projects is best handled by session cast merging, which integrates personal agents alongside project agents so they participate in the same workflow.

Personal agent discovery merges personal agents alongside project agents to apply consistent advising behavior across multi-project workflows while strictly avoiding the modification of project-managed state.