mcpField
T
team-init

team-init

github.com/tikalkAgent workflows143137 installsUpdated today
Repository

About team-init

team-init is an agent skill for examining existing codebases and documenting reusable conventions as Context Directive Record (CDR) drafts. It is intended for brownfield projects, legacy modernization, and team onboarding.

Key features

  1. 01Detects codebase subsystems
  2. 02Finds reusable implementation patterns
  3. 03Compares against team directives
  4. 04Writes CDR drafts to adlc branch
  5. 05Hands off drafts for review

Use cases

  1. 01Document conventions in an existing project
  2. 02Capture patterns before legacy refactoring
  3. 03Make implicit practices explicit for onboarding

Install

Copy the command and run it in a terminal.

$npx skills add https://github.com/tikalk/adlc-team-skills --skill team-init

After installing, type/team-initin Claude Code to confirm it loaded.

Bundled filesscripts/bash/team-init.shscripts/powershell/team-init.ps1

Details

From the project README

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-levelup to extract from session
  • Mining git history: Use /change-init for past decisions

Process

Phase 0: Environment Setup

Run: scripts/bash/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-levelup 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

Comments

Markdown supported