Initialize Asha in Project

Sets up Asha framework for the current project.

Arguments: $ARGUMENTS

What This Creates

${CLAUDE_PROJECT_DIR}/
├── Memory/
│   ├── sessions/archive/
│   ├── reasoning_bank/
│   ├── vector_db/
│   ├── activeContext.md
│   ├── projectbrief.md
│   ├── communicationStyle.md
│   ├── workflowProtocols.md
│   └── techEnvironment.md
├── Work/markers/
├── .asha/
│   ├── .venv/
│   └── config.json
└── CLAUDE.md

Protocol

Step 1: Check Existing Installation

if [[ -f "${CLAUDE_PROJECT_DIR}/.asha/config.json" ]]; then
    echo "Asha already initialized in this project"
    echo "To reinitialize, delete .asha/ and run again"
    exit 0
fi

If already initialized, inform user and stop.

Step 2: Create Directory Structure

mkdir -p "${CLAUDE_PROJECT_DIR}/Memory/sessions/archive"
mkdir -p "${CLAUDE_PROJECT_DIR}/Memory/reasoning_bank"
mkdir -p "${CLAUDE_PROJECT_DIR}/Memory/vector_db"
mkdir -p "${CLAUDE_PROJECT_DIR}/Work/markers"
mkdir -p "${CLAUDE_PROJECT_DIR}/.asha"

Step 3: Copy Templates (if Memory files don't exist)

For each template in ${CLAUDE_PLUGIN_ROOT}/templates/:

• If Memory/<filename> doesn't exist, copy it
• If it exists, skip (preserve user content)

Templates to copy:

• activeContext.md
• projectbrief.md
• communicationStyle.md
• workflowProtocols.md
• techEnvironment.md

for template in activeContext.md projectbrief.md communicationStyle.md workflowProtocols.md techEnvironment.md; do
    if [[ ! -f "${CLAUDE_PROJECT_DIR}/Memory/$template" ]]; then
        cp "${CLAUDE_PLUGIN_ROOT}/templates/$template" "${CLAUDE_PROJECT_DIR}/Memory/$template"
        echo "Created Memory/$template"
    else
        echo "Skipped Memory/$template (exists)"
    fi
done

Step 4: Create CLAUDE.md (if doesn't exist)

if [[ ! -f "${CLAUDE_PROJECT_DIR}/CLAUDE.md" ]]; then
    cp "${CLAUDE_PLUGIN_ROOT}/templates/CLAUDE.md" "${CLAUDE_PROJECT_DIR}/CLAUDE.md"
    echo "Created CLAUDE.md"
else
    echo "Skipped CLAUDE.md (exists)"
fi

Step 5: Create Python Virtual Environment

Unless --minimal is specified:

python3 -m venv "${CLAUDE_PROJECT_DIR}/.asha/.venv"
"${CLAUDE_PROJECT_DIR}/.asha/.venv/bin/pip" install -r "${CLAUDE_PLUGIN_ROOT}/tools/requirements.txt"

If venv creation fails, warn but continue (Vector DB will be unavailable).

Step 6: Initialize Databases

# ReasoningBank
"${CLAUDE_PLUGIN_ROOT}/tools/run-python.sh" "${CLAUDE_PLUGIN_ROOT}/tools/reasoning_bank.py" stats

# Vector DB check
"${CLAUDE_PLUGIN_ROOT}/tools/run-python.sh" "${CLAUDE_PLUGIN_ROOT}/tools/memory_index.py" check

Step 7: Create Config File

Write .asha/config.json to mark project as initialized:

cat > "${CLAUDE_PROJECT_DIR}/.asha/config.json" << 'EOF'
{
  "version": "1.0.0",
  "initialized": "$(date -Iseconds)",
  "plugin": "asha@asha-marketplace"
}
EOF

Step 8: Report Status

Display:

• Directory structure created
• Templates copied (list which ones)
• Python venv status
• Vector DB readiness
• Next steps for user

Next Steps After Init

1. Edit Memory/projectbrief.md - Define project scope
2. Edit Memory/activeContext.md - Set current status
3. Run /asha:index - Index files for semantic search (optional)
4. Add to .gitignore:

   .asha/
   Memory/sessions/
   Memory/vector_db/
   Memory/reasoning_bank/