📄️ Agent Anatomy
The four-component architecture — Brain, Skills, Tools, Guardrails — and why the separation matters for governance and operational reliability
📄️ SKILL.md Authoring Guide
Why skills beat prompts, the two-zone design (Scripts Zone + Agents Zone), RUBRIC.md four-tier quality framework, and the skill lifecycle
📄️ Tool Integration Patterns
CLI vs MCP vs mock wrapper patterns, DANGEROUS_PATTERNS mechanical safety gate, approval modes, and fleet delegation tools
📄️ Governance Reference
Two-layer governance model, L1-L4 maturity levels, approval modes, DANGEROUS_PATTERNS categories, audit logging, and promotion framework
📄️ Profile Guide
Profile = agent definition, SOUL.md anatomy, config.yaml anatomy, skills/ subdirectory, fleet coordinator pattern, and profile completeness checklist