team-init

What this skill does

Reverse-engineer Context Directive Records (CDRs) from an existing codebase (brownfield) to document reusable patterns for contribution to team-ai-directives.

Replaces the former team-init skill. Drafts are written to the adlc orphan branch of team-ai-directives (not project-local .adlc/drafts/).

When to use

• Brownfield projects: Existing code without team-wide directives
• Legacy modernization: Extract reusable patterns before refactoring
• Team onboarding: Turn implicit conventions into explicit directives

When NOT to use

• Greenfield (session complete): Use /team-learn to extract from session
• Mining git history: Use /change-init for past decisions

Process

Phase 0: Environment Setup

Run: scripts/team-init.sh --setup

Resolves REPO_ROOT, TEAM_AI_DIRECTIVES, NEXT_CDR, ADLC_BRANCH.

Phase 1: Sub-System Detection

Analyze codebase for distinct sub-systems (same detection as former team-init):

• src/auth/ → Authentication
• services/payment/ → Payment
• etc.

Phase 2: Load Team Directives

Read existing team-ai-directives main branch for comparison (avoid duplicates).

Phase 3: Multi-Agent Analysis

1. Discovery Agent: Scan each sub-system for raw patterns
2. Pattern Agent: Classify and score patterns for reusability
3. Synthesis Agent: Cross-sub-system analysis and CDR generation

Phase 4: Write CDRs

Write to adlc branch drafts/cdr/ using the shared CDR draft template at skills/team/templates/cdr-draft-template.md.

Phase 5: Handoff

Suggest running /team-learn to review and publish the discovered CDRs.

Verification

• CDRs written to adlc branch drafts/cdr/
• No CDRs in .adlc/drafts/ (old location)
• cdr.md index regenerated in adlc branch