implement-epic

/implement-epic - Parallel Sprint Execution

Workflow position: epic → plan → tasks → implement-epic → optimize → ship

Key principles:

• Layers execute sequentially (dependencies respected)
• Sprints within a layer run in parallel (single message, multiple Task calls)
• Workers implement one feature at a time (Domain Memory pattern)
• Errors classified and handled via error-recovery skill

Epic Directory: !ls -td epics/[0-9]*-* 2>/dev/null | head -1

Sprint Plan: !cat epics/*/sprint-plan.md 2>/dev/null | head -30

Sprint Count: !grep -c "^## Sprint S" epics/*/sprint-plan.md 2>/dev/null || echo "0"

Execution Layers: !bash .spec-flow/scripts/bash/sprint-utils.sh layers "$(ls -td epics/[0-9]*-* 2>/dev/null | head -1)" 2>/dev/null

Auto Mode: !echo "$ARGUMENTS" | grep -q "\-\-auto-mode" && echo "enabled" || echo "disabled"

Step 1: Validate Prerequisites

EPIC_DIR=$(ls -td epics/[0-9]*-* 2>/dev/null | head -1)

if [ -z "$EPIC_DIR" ]; then
    echo "ERROR: No epic directory found"
    echo "Run /epic first to create an epic"
    exit 1
fi

# Validate sprint directories exist
bash .spec-flow/scripts/bash/sprint-utils.sh validate "$EPIC_DIR"

Read sprint-plan.md to understand execution layers:

cat "$EPIC_DIR/sprint-plan.md"

Step 2: Initialize Domain Memory (If Missing)

For each sprint without domain-memory.yaml, initialize it:

for sprint_dir in "$EPIC_DIR/sprints"/*/; do
    if [ ! -f "${sprint_dir}/domain-memory.yaml" ]; then
        SPRINT_ID=$(basename "$sprint_dir")
        echo "Initializing domain memory for $SPRINT_ID..."

        # Use initializer agent
        Task(subagent_type="initializer", prompt="
            Initialize domain memory for sprint.
            Sprint directory: ${sprint_dir}
            Tasks file: ${sprint_dir}/tasks.md
            Workflow type: epic-sprint
        ")
    fi
done