Alex Newman
8d485890b9
- Introduced functionality for installing, uninstalling, and checking the status of Cursor hooks. - Added a new command structure for managing hooks with detailed usage instructions. - Implemented a method to locate the cursor-hooks directory across different environments. - Updated build-hooks script to inform users about the location of Cursor hooks. This enhancement streamlines the integration of Claude-Mem with Cursor, improving user experience and accessibility of hooks.
112 lines
2.9 KiB
Bash
Executable File
112 lines
2.9 KiB
Bash
Executable File
#!/bin/bash
|
|
# Session Summary Hook for Cursor (stop)
|
|
# Called when agent loop ends
|
|
#
|
|
# This hook:
|
|
# 1. Generates session summary
|
|
# 2. Updates context file for next session
|
|
#
|
|
# Output: Empty JSON {} or {"followup_message": "..."} for auto-iteration
|
|
|
|
# Source common utilities
|
|
SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
|
|
source "${SCRIPT_DIR}/common.sh" 2>/dev/null || {
|
|
echo '{}'
|
|
exit 0
|
|
}
|
|
|
|
# Check dependencies (non-blocking)
|
|
check_dependencies >/dev/null 2>&1 || true
|
|
|
|
# Read JSON input from stdin with error handling
|
|
input=$(read_json_input)
|
|
|
|
# Extract common fields with safe fallbacks
|
|
conversation_id=$(json_get "$input" "conversation_id" "")
|
|
generation_id=$(json_get "$input" "generation_id" "")
|
|
workspace_root=$(json_get "$input" "workspace_roots[0]" "")
|
|
status=$(json_get "$input" "status" "completed")
|
|
|
|
# Fallback workspace to current directory
|
|
if is_empty "$workspace_root"; then
|
|
workspace_root=$(pwd)
|
|
fi
|
|
|
|
# Get project name
|
|
project_name=$(get_project_name "$workspace_root")
|
|
|
|
# Use conversation_id as session_id, fallback to generation_id
|
|
session_id="$conversation_id"
|
|
if is_empty "$session_id"; then
|
|
session_id="$generation_id"
|
|
fi
|
|
|
|
# Exit if no session_id available
|
|
if is_empty "$session_id"; then
|
|
echo '{}'
|
|
exit 0
|
|
fi
|
|
|
|
# Get worker port from settings with validation
|
|
worker_port=$(get_worker_port)
|
|
|
|
# Ensure worker is running (with retries)
|
|
if ! ensure_worker_running "$worker_port"; then
|
|
echo '{}'
|
|
exit 0
|
|
fi
|
|
|
|
# 1. Request summary generation (fire-and-forget)
|
|
# Note: Cursor doesn't provide transcript_path like Claude Code does,
|
|
# so we can't extract last_user_message and last_assistant_message.
|
|
payload=$(jq -n \
|
|
--arg sessionId "$session_id" \
|
|
'{
|
|
contentSessionId: $sessionId,
|
|
last_user_message: "",
|
|
last_assistant_message: ""
|
|
}' 2>/dev/null)
|
|
|
|
if [ -n "$payload" ]; then
|
|
curl -s -X POST \
|
|
"http://127.0.0.1:${worker_port}/api/sessions/summarize" \
|
|
-H "Content-Type: application/json" \
|
|
-d "$payload" \
|
|
>/dev/null 2>&1 &
|
|
fi
|
|
|
|
# 2. Update context file for next session
|
|
# Fetch fresh context (includes observations from this session)
|
|
project_encoded=$(url_encode "$project_name")
|
|
context=$(curl -s -f "http://127.0.0.1:${worker_port}/api/context/inject?project=${project_encoded}" 2>/dev/null || echo "")
|
|
|
|
if [ -n "$context" ]; then
|
|
rules_dir="${workspace_root}/.cursor/rules"
|
|
rules_file="${rules_dir}/claude-mem-context.mdc"
|
|
|
|
# Create rules directory if it doesn't exist
|
|
mkdir -p "$rules_dir" 2>/dev/null || true
|
|
|
|
# Write context as a Cursor rule with alwaysApply: true
|
|
cat > "$rules_file" 2>/dev/null << EOF
|
|
---
|
|
alwaysApply: true
|
|
description: "Claude-mem context from past sessions (auto-updated)"
|
|
---
|
|
|
|
# Memory Context from Past Sessions
|
|
|
|
The following context is from claude-mem, a persistent memory system that tracks your coding sessions.
|
|
|
|
${context}
|
|
|
|
---
|
|
*Updated after last session. Use claude-mem's MCP search tools for more detailed queries.*
|
|
EOF
|
|
fi
|
|
|
|
# Output empty JSON - no followup message
|
|
echo '{}'
|
|
exit 0
|
|
|