What problem does it solve? Creating an APX project agent without a system prompt silently produces a useless agent, and hand-editing agent files leaves the running daemon unaware of changes. This Skill guides the correct creation, configuration, renaming, and memory management of project agents so they actually work. ## Core Features & Use Cases - Agent creation with prompts: Create agents via the create_agent tool or apx agent add --prompt -, always including the body that becomes the agent's system prompt. - Identity configuration: Set typology (orchestrator, specialist, assistant, worker, monitor), org-chart area/role slugs, avatar blob icons, and per-agent models. - Lifecycle management: Safely rename agents with rename_agent (repointing all slug references), seed per-agent memory with write_agent_memory, and import agents from the global vault. - Use Case: You need a code reviewer agent for your project. Run apx agent add reviewer --type specialist --area engineering --prompt - with a heredoc containing its responsibilities and hard limits, then verify with apx agent get reviewer. ## Quick Start Ask the assistant to create a new APX project agent with a specific role, area, and a full system prompt describing its responsibilities and limits.
