ai-first-engineeringlisted

# CIEL ADAPTATION: AI-First Engineering This skill provides CIEL's core philosophy for engineering in an AI-native environment. It prioritizes planning quality, deterministic verification, and agent-friendly architectures. ## Integration Context Adapted from `~/.agents/skills/ai-first-engineering/`. This is CIEL's "Cultural Directive" for how it designs its own processes and reviews. ## Core Mandates ### 1. Planning Over Typing The **orchestration** skill must prioritize the "Research -> Strategy" phase. Implementation is a derivative of a well-defined plan. ### 2. Agent-Friendly Architecture When proposing refactors or new features, prioritize: - **Explicit Boundaries**: Decoupled services and clear API contracts. - **Typed Interfaces**: Use TypeScript, Zod, or formal schemas for all data flow. - **Deterministic Tests**: No flaky tests; every failure must be reproducible. ### 3. Elevated Testing Standard For all CIEL-orchestrated work: - Regression coverage is mandatory for touched domains. - Explicit assertions for edge cases and failure modes. - Integration checks for all interface boundaries. ## Orchestration Logic ### 1. Architectural Review The **Council of Five** must evaluate all major architectural proposals against the "Agent-Friendly" criteria. ### 2. Review Focus CIEL's internal code review (via **agentic-engineering**) must focus on system behavior, security, and integrity, ignoring style issues covered by automation. ### 3. Hiring & Evolution