Skip to content

onboarding.md

Latest commit

 

History

History
37 lines (30 loc) · 1.9 KB

onboarding.md

File metadata and controls

37 lines (30 loc) · 1.9 KB
Raw

SYSTEM INITIALIZATION: AGENTIC WORKFLOW BOOTSTRAP

You are entering a structured, transparent engineering environment.

1. IDENTITY & REASONING ENGINE

  • Primary Skill: Load .ai/skills/Engineer.md.
  • Think Out Loud: You are required to provide a visible Chain-of-Thought. Before executing any code, walk through your methodology, logic, and potential hardware implications in the chat.

2. STATE SYNCHRONIZATION

Before implementation:

  1. Index Project: Read .ai/specs/requirements.md, design.md, and tasks.md.
  2. Locate Pointer: Identify the active task in tasks.md.

3. NATIVE CONFIGURATION

  • Generate/Update your native intelligence file (e.g., CLAUDE.md, .cursorrules).
  • Instruction: "Always perform a visible step-by-step breakdown in chat before implementation. Mirror all progress in the .ai/ directory."

4. INITIAL SYNC REPORT

Provide a detailed response:

  • Status: Which task are we starting?
  • Methodology: Your "Thinking Out Loud" analysis of how to approach this task.
  • Sub-tasks: A breakdown of mini-steps you will follow.

5. ENVIRONMENT DETECTION

Identify your environment and load the corresponding native configuration rule:

  • Claude Code CLI: Target CLAUDE.md
  • Cursor IDE: Target .cursorrules
  • GitHub Copilot: Target .github/copilot-instructions.md
  • Generic/Web: Target .ai/SESSION_LOG.md

6. THE SYNC COMMAND

  1. Load Identity: Read .ai/skills/Engineer.md.
  2. Scan State: Read all files in .ai/specs/.
  3. Generate/Sync: If the native file for your environment (from Step 1) is missing or outdated compared to the .ai/ directory, update it immediately.

7. DATA PROPAGATION

If new project data or a new "Spec" is created during this session, you are mandated to propagate that change to the native config file so future sessions are "born" with that knowledge. Awaiting your analysis before we begin.