Index Memory for Semantic Search

Run the memory index tool to enable semantic search across project files.

Arguments: $ARGUMENTS

Protocol

Step 1: Verify Initialization

Check that the project has been initialized with Asha:

[[ -f "${CLAUDE_PROJECT_DIR}/.asha/config.json" ]] && echo "Asha initialized" || echo "ERROR: Run /asha:init first"

If not initialized, instruct user to run /asha:init.

Step 2: Determine Mode

Based on arguments:

• --full → Full reindex of all files
• --check → Verify dependencies only (Ollama running, packages installed)
• No arguments → Incremental update (changed files only, faster)

Step 3: Run Indexer

Use the run-python.sh wrapper which auto-detects the virtual environment:

# Incremental (default - changed files only)
"${CLAUDE_PLUGIN_ROOT}/tools/run-python.sh" "${CLAUDE_PLUGIN_ROOT}/tools/memory_index.py" ingest --changed

# Full reindex
"${CLAUDE_PLUGIN_ROOT}/tools/run-python.sh" "${CLAUDE_PLUGIN_ROOT}/tools/memory_index.py" ingest

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

Step 4: Report Results

Summarize:

• Number of files indexed and chunks created
• Any errors or warnings encountered
• If dependencies are missing, provide install instructions

Requirements

• Ollama running locally (ollama serve)
• Embedding model available (ollama pull nomic-embed-text)
• Python packages installed (run /asha:init to set up)