Sbrain turns agent skills into loops

Sbrain’s interesting bet is that reliable agent autonomy comes from operating structure, not larger prompts.

• –CEO and Cassandra skills create initiative and independent dissent as complementary functions
• –Durable evidence, reporting cadences, and authority boundaries make agents more accountable over time
• –The skill-cutter keeps these systems from becoming bloated context dumps
• –The experimental skill-sync CLI points toward portable agent organizations across projects and machines
• –It remains very alpha, so the real test is whether these loops produce useful decisions rather than procedural theater