What problem does it solve? When a new AI agent or a new session joins a multi-agent project, it lacks context about the repository state, past decisions, active sprint goals, and forbidden approaches. This Skill automates context synchronization so agents start work with an accurate, shared understanding instead of guessing. ## Core Features & Use Cases - Git State Inspection: Checks the current branch, uncommitted changes, recent commits, and active Git worktrees to assess repository readiness. - SSOT Memory Ingestion: Reads .agents/MEMORY.md, task.md, and .agents/plans/ to surface architecture decisions, sprint goals, and explicitly rejected approaches. - Executive Briefing Output: Produces a structured 5-section report covering repository state, persistent memory, current sprint, implementation plans, and governance rules. - Use Case: A Claude Code session starts on a project managed by AGENTS-OS. Running the onboarding script instantly reveals the branch is behind origin, summarizes the last five commits, lists the active sprint goal, and reminds the agent never to push directly to main. ## Quick Start Run the swarm-onboarding skill to scan the project and generate an executive briefing of the repository state, memory, and current sprint before starting any coding task.
