arc-using-worktrees

When to Use

Use this skill when you need to create an isolated git worktree for developing an epic. This enables parallel development of multiple epics without interference.

IMPORTANT: This skill is specific to arcforge. Do NOT use any other worktree skill. The metadata format is different.

Core Workflow

Step 1: Verify Directory Structure

Check if .worktrees/ directory exists:

ls .worktrees/ 2>/dev/null || mkdir -p .worktrees/

Step 1.5: Safety Verification

Ensure .worktrees/ is in .gitignore:

git check-ignore -q .worktrees 2>/dev/null
if [ $? -ne 0 ]; then
    echo ".worktrees" >> .gitignore
    git add .gitignore
    git commit -m "chore: ignore .worktrees directory"
fi

Step 1.6: Worktree Context Check

If already in a worktree, STOP and use current worktree:

git rev-parse --show-superproject-working-tree >/dev/null 2>&1
if [ $? -eq 0 ]; then
    echo "Already in a worktree. Stay here and continue."
    exit 1
fi

Step 2: Create Git Worktree

Create the worktree with a matching branch name:

git worktree add .worktrees/<epic-name> -b <epic-name>

Important: Use the exact epic name from dag.yaml if it exists.

Step 2.5: Project Setup

Auto-detect and run setup:

# Python
if [ -f pyproject.toml ]; then pip install -e .; fi
if [ -f requirements.txt ]; then pip install -r requirements.txt; fi

# Node
if [ -f package.json ]; then npm install; fi

Step 2.6: Baseline Test

Verify clean baseline:

pytest tests/ -v

If tests fail: Report failures, ask whether to proceed.

Step 3: Create Epic Metadata

CRITICAL: Create .arcforge-epic file (NOT .epic, NOT .worktree, NOT any other name):

cd .worktrees/<epic-name>
echo "<epic-name>" > .arcforge-epic

Why .arcforge-epic specifically?

• •Follows project naming convention (.arcforge-progress.json, etc.)
• •Ensures arc-coordinating CLI can find it
• •Distinguishes from other project's worktree tracking

DO NOT:

• •❌ Create .epic (wrong name)
• •❌ Create .worktree (wrong name)
• •❌ Skip this file (required for tracking)
• •❌ Use other tracking mechanisms (this is the standard)

Step 4: DAG Integration (if dag.yaml exists)

If dag.yaml exists in the main worktree:

• •Read the epic status
• •Verify the epic is in "ready" state (dependencies complete)
• •Note: Status updates happen via arc-coordinating, not here

Step 5: Notify User

Use the standardized completion format (see below).

Red Flags

Stop immediately if you catch yourself thinking:

1. •".epic is fine" — NO! Must use .arcforge-epic exactly
2. •"I'll use another worktree skill" — NO! Wrong metadata format
3. •"Skip DAG check" — Always check if dag.yaml exists
4. •"Create worktree anywhere" — Must be in .worktrees/ directory
5. •"Use main/master branch" — Must create new branch with epic name
6. •"Metadata file is optional" — It's REQUIRED for coordinator tracking
7. •"Any name starting with dot is fine" — NO! .arcforge-epic ONLY

Common Rationalizations

Stage Completion Format

─────────────────────────────────────────────────
✅ Worktree created → `.worktrees/<epic-name>/`

Branch: <epic-name>
Tracking: .arcforge-epic
DAG status: [ready/blocked/N/A]

Next: Work in the worktree, then use `/arc-finishing-epic` when complete
─────────────────────────────────────────────────

Blocked Format

─────────────────────────────────────────────────
⚠️ Worktree creation blocked

Issue: [description]
Location: [epic name or path]

To resolve:
1. [action]

Then retry: `/arc-using-worktrees`
─────────────────────────────────────────────────

Integration with Pipeline

• •Called by: arc-agent-driven, arc-executing-tasks, arc-coordinating
• •Before: arc-coordinating expand (creates multiple worktrees)
• •After: Work in worktree, then /arc-finishing-epic (merge decision)
• •Related: arc-dispatching-parallel (for feature-level parallelization within worktree)