claude-flow-novice 2.18.6 → 2.18.7

package/claude-assets/cfn-extras/skills/utility/cfn-process-instrumentation/SKILL.md

@@ -1,279 +0,0 @@
-# CFN Process Instrumentation Skill
-
-## Purpose
-
-Provides comprehensive process instrumentation, monitoring, and automatic resource limiting for CFN Loop agents and orchestration processes. This skill is essential for detecting and preventing memory leaks before they impact system performance.
-
-## Core Functions
-
-### Process Monitoring
-- Real-time memory usage tracking
-- CPU utilization monitoring
-- File handle and thread tracking
-- Automatic resource limit enforcement
-
-### Telemetry Collection
-- Structured metrics collection in JSON format
-- Time-series data for performance analysis
-- Process lifecycle tracking
-- Exit code and duration recording
-
-### Resource Limiting
-- Memory limit enforcement with automatic termination
-- CPU usage throttling and monitoring
-- Timeout protection for long-running processes
-- Configurable limits per agent type
-
-## Usage Patterns
-
-### Automatic Instrumentation
-```bash
-#!/usr/bin/env bash
-
-# Source to automatically instrument current process
-source "./cfn-process-instrumentation/instrument-process.sh"
-
-# Rest of script runs with monitoring
-```
-
-### Manual Process Monitoring
-```bash
-# Monitor specific PID
-./cfn-process-instrumentation/instrument-process.sh monitor-pid 12345
-
-# Instrument with custom limits
-AGENT_ID="my-agent" CFN_MEMORY_LIMIT="1G" \
-    ./instrument-process.sh
-```
-
-### Integration in Agent Spawning
-```bash
-#!/usr/bin/env bash
-
-# Load instrumentation before spawning agent
-source "./cfn-process-instrumentation/instrument-process.sh"
-
-# Spawn agent with monitoring
-npx claude-flow-novice agent "$AGENT_TYPE" \
-    --max-memory "$CFN_MEMORY_LIMIT" \
-    --timeout "$CFN_TIMEOUT"
-```
-
-## Configuration
-
-### Environment Variables
-- `AGENT_ID`: Unique identifier for the agent process
-- `CFN_MEMORY_LIMIT`: Memory limit (default: 2G, formats: 2G, 2048M, etc.)
-- `CFN_CPU_LIMIT`: CPU usage limit as percentage (default: 80%)
-- `CFN_TIMEOUT`: Maximum execution time in seconds (default: 600)
-- `CFN_TELEMETRY_DIR`: Directory for metrics storage (default: /tmp/cfn-telemetry)
-
-### Resource Limits
-The skill enforces configurable limits:
-- **Memory**: Prevents OOM conditions with automatic termination
-- **CPU**: Throttles excessive CPU usage
-- **Timeout**: Kills processes that exceed time limits
-- **File Handles**: Monitors for file descriptor leaks
-
-## Telemetry Data Structure
-
-### Metrics Format
-```json
-{
-  "agent_id": "agent-123",
-  "start_time": "2025-01-06T10:00:00Z",
-  "process_id": "12345",
-  "memory_limit": "2G",
-  "cpu_limit": "80%",
-  "timeout": "600",
-  "end_time": "2025-01-06T10:05:00Z",
-  "exit_code": "0",
-  "samples": [
-    {
-      "timestamp": "2025-01-06T10:01:00Z",
-      "memory_kb": "102400",
-      "cpu_percent": "15.5",
-      "open_files": "45",
-      "threads": "8"
-    }
-  ]
-}
-```
-
-### Collected Metrics
-- **Memory Usage**: RSS memory in kilobytes
-- **CPU Usage**: Percentage of CPU utilization
-- **Open Files**: Number of open file descriptors
-- **Threads**: Number of active threads
-- **Timestamp**: ISO 8601 formatted time
-
-## Integration Points
-
-### Orchestration Scripts
-```bash
-#!/usr/bin/env bash
-
-# Load instrumentation for orchestrator
-source "./cfn-process-instrumentation/instrument-process.sh"
-
-# Orchestrator runs with monitoring
-echo "Starting orchestration with process monitoring"
-```
-
-### Agent Templates
-```bash
-#!/usr/bin/env bash
-
-# All agents get automatic instrumentation
-source "./cfn-process-instrumentation/instrument-process.sh"
-
-# Agent execution code
-echo "Agent execution monitored and resource-limited"
-```
-
-### CFN Loop Integration
-- **Loop 1**: Orchestration process monitoring
-- **Loop 2**: Validator process tracking
-- **Loop 3**: Implementation agent monitoring
-- **Loop 4**: Product Owner process tracking
-
-## Safety Features
-
-### Graceful Shutdown
-- SIGTERM sent before SIGKILL
-- Cleanup traps for proper resource release
-- Final report generation on exit
-
-### Error Handling
-- Missing tools handled gracefully
-- Fallback behavior when utilities unavailable
-- Comprehensive error logging
-
-### Resource Protection
-- Automatic termination on limit violation
-- Background monitoring with minimal overhead
-- Process group management for cleanup
-
-## Performance Considerations
-
-### Monitoring Overhead
-- Minimal impact on agent performance (<1% CPU)
-- Efficient metric collection with native tools
-- Configurable monitoring intervals (default: 30s)
-
-### Storage Efficiency
-- JSON metrics with compact storage
-- Automatic cleanup of old telemetry files
-- Configurable retention policies
-
-### Network Independence
-- No external dependencies for monitoring
-- Local process inspection only
-- Works in isolated environments
-
-## Usage Examples
-
-### Basic Agent Monitoring
-```bash
-#!/usr/bin/env bash
-
-# Simple agent with monitoring
-source "./cfn-process-instrumentation/instrument-process.sh"
-
-# Agent work here
-echo "Performing agent tasks..."
-
-# Monitoring and cleanup automatic
-```
-
-### Custom Resource Limits
-```bash
-#!/usr/bin/env bash
-
-# Agent with custom limits
-export CFN_MEMORY_LIMIT="1G"
-export CFN_CPU_LIMIT="60%"
-export CFN_TIMEOUT="300"
-
-source "./cfn-process-instrumentation/instrument-process.sh"
-
-# Agent runs with stricter limits
-```
-
-### Background Process Monitoring
-```bash
-#!/usr/bin/env bash
-
-# Start long-running process
-./long-running-task &
-TASK_PID=$!
-
-# Monitor the background process
-./cfn-process-instrumentation/instrument-process.sh monitor-pid $TASK_PID
-
-# Wait for completion
-wait $TASK_PID
-```
-
-## Monitoring and Alerting
-
-### Threshold Alerts
-- Memory usage approaching limits
-- CPU usage sustained at high levels
-- File descriptor count growing unexpectedly
-- Process hanging without activity
-
-### Telemetry Analysis
-```bash
-# Analyze metrics for trends
-jq '.samples | map(.memory_kb) | add / length' /tmp/cfn-telemetry/metrics_*.json
-
-# Find processes with high memory usage
-find /tmp/cfn-telemetry -name "*.json" -exec jq 'if (.samples | length) > 0 then (.samples[-1].memory_kb | tonumber) else 0 end' {} \;
-```
-
-### Integration with Monitoring Systems
-- JSON metrics compatible with log aggregators
-- Time-series data for graphing tools
-- Structured alerts for DevOps monitoring
-
-## Troubleshooting
-
-### Common Issues
-1. **High monitoring overhead**: Increase monitoring interval
-2. **Missing tools**: Install ps, lsof, jq for full functionality
-3. **Permission denied**: Check file system permissions for telemetry directory
-4. **Process termination**: Review and adjust resource limits
-
-### Debug Mode
-```bash
-# Enable detailed logging
-export CFN_DEBUG=1
-source "./cfn-process-instrumentation/instrument-process.sh"
-```
-
-### Manual Inspection
-```bash
-# Check current metrics
-cat /tmp/cfn-telemetry/metrics_*.json | jq .
-
-# Monitor process manually
-ps -p $$ -o pid,rss,pcpu,nlwp,fd
-```
-
-## Security Considerations
-
-### Data Protection
-- Telemetry stored in local filesystem only
-- No network transmission of metrics
-- Configurable storage location
-
-### Process Isolation
-- Each agent tracked independently
-- No cross-agent data sharing
-- Clean separation of metrics
-
-### Access Control
-- File permissions restricted to agent user
-- No privileged operations required
-- Safe execution in untrusted environments

package/claude-assets/cfn-extras/skills/utility/cfn-process-instrumentation/instrument-process.sh

@@ -1,327 +0,0 @@
-#!/usr/bin/env bash
-
-##############################################################################
-# CFN Process Instrumentation
-# Part of ANTI-023 Memory Leak Protection System
-#
-# Provides process instrumentation, monitoring, and automatic resource limiting
-# for CFN Loop agents and orchestration processes.
-#
-# Usage:
-#   source ./instrument-process.sh [--agent-id <id>] [--memory-limit <size>]
-#   ./instrument-process.sh --monitor-pid <pid>
-##############################################################################
-
-set -euo pipefail
-
-# Configuration
-SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
-PROJECT_ROOT="$(cd "$SCRIPT_DIR/../.." && pwd)"
-
-# Default limits
-# Note: DEFAULT_TIMEOUT may be set by wrapped-executor.sh (300)
-# Only set if not already defined
-: "${DEFAULT_MEMORY_LIMIT:="2G"}"
-: "${DEFAULT_CPU_LIMIT:="80%"}"
-: "${DEFAULT_TIMEOUT:="600"}"
-
-# Process tracking
-AGENT_ID="${AGENT_ID:-$(hostname)-$$}"
-MONITOR_PID=""
-MEMORY_LIMIT="${CFN_MEMORY_LIMIT:-$DEFAULT_MEMORY_LIMIT}"
-CPU_LIMIT="${CFN_CPU_LIMIT:-$DEFAULT_CPU_LIMIT}"
-TIMEOUT="${CFN_TIMEOUT:-$DEFAULT_TIMEOUT}"
-
-# Telemetry storage
-TELEMETRY_DIR="${CFN_TELEMETRY_DIR:-/tmp/cfn-telemetry}"
-METRICS_FILE="$TELEMETRY_DIR/metrics_${AGENT_ID}.json"
-
-# Color coding
-readonly RED='\033[0;31m'
-readonly GREEN='\033[0;32m'
-readonly YELLOW='\033[1;33m'
-readonly BLUE='\033[0;34m'
-readonly NC='\033[0m'
-
-# Logging functions
-log_info() {
-    echo -e "${BLUE}[INSTRUMENT]${NC} $1" >&2
-}
-
-log_success() {
-    echo -e "${GREEN}[INSTRUMENT]${NC} $1" >&2
-}
-
-log_warning() {
-    echo -e "${YELLOW}[INSTRUMENT]${NC} $1" >&2
-}
-
-log_error() {
-    echo -e "${RED}[INSTRUMENT]${NC} $1" >&2
-}
-
-# Initialize telemetry directory
-init_telemetry() {
-    mkdir -p "$TELEMETRY_DIR"
-
-    # Create metrics file with initial structure
-    cat > "$METRICS_FILE" << EOF
-{
-  "agent_id": "$AGENT_ID",
-  "start_time": "$(date -Iseconds)",
-  "process_id": "$$",
-  "memory_limit": "$MEMORY_LIMIT",
-  "cpu_limit": "$CPU_LIMIT",
-  "timeout": "$TIMEOUT",
-  "samples": []
-}
-EOF
-}
-
-# Collect process metrics
-collect_metrics() {
-    local pid="${1:-$$}"
-    local timestamp="$(date -Iseconds)"
-
-    # Get process statistics
-    local mem_usage=""
-    local cpu_usage=""
-    local open_files=""
-    local threads=""
-
-    if command -v ps >/dev/null 2>&1; then
-        mem_usage=$(ps -p "$pid" -o rss= 2>/dev/null | tr -d ' ' || echo "0")
-        cpu_usage=$(ps -p "$pid" -o %cpu= 2>/dev/null | tr -d ' ' || echo "0")
-    fi
-
-    if command -v lsof >/dev/null 2>&1; then
-        open_files=$(lsof -p "$pid" 2>/dev/null | wc -l || echo "0")
-    fi
-
-    if [[ -f "/proc/$pid/status" ]]; then
-        threads=$(grep "^Threads:" "/proc/$pid/status" | awk '{print $2}' || echo "0")
-    fi
-
-    # Create metrics entry
-    local metrics_entry=$(cat << EOF
-{
-  "timestamp": "$timestamp",
-  "memory_kb": "$mem_usage",
-  "cpu_percent": "$cpu_usage",
-  "open_files": "$open_files",
-  "threads": "$threads"
-}
-EOF
-    )
-
-    # Update metrics file
-    if [[ -f "$METRICS_FILE" ]]; then
-        # Use jq to safely append to samples array
-        if command -v jq >/dev/null 2>&1; then
-            jq --argjson entry "$metrics_entry" '.samples += [$entry]' "$METRICS_FILE" > "$METRICS_FILE.tmp" && \
-                mv "$METRICS_FILE.tmp" "$METRICS_FILE"
-        else
-            # Fallback without jq
-            echo "Warning: jq not available, using simple append" >&2
-            echo "$metrics_entry" >> "$METRICS_FILE.raw"
-        fi
-    fi
-}
-
-# Check resource limits
-check_limits() {
-    local pid="${1:-$$}"
-
-    # Memory limit check
-    if command -v ps >/dev/null 2>&1; then
-        local mem_kb=$(ps -p "$pid" -o rss= 2>/dev/null | tr -d ' ' || echo "0")
-        local mem_mb=$((mem_kb / 1024))
-
-        case "$MEMORY_LIMIT" in
-            *G|*g)
-                local limit_mb=$((${MEMORY_LIMIT%[Gg]*} * 1024))
-                ;;
-            *M|*m)
-                local limit_mb=$((${MEMORY_LIMIT%[Mm]*}))
-                ;;
-            *)
-                local limit_mb=2048  # Default 2GB
-                ;;
-        esac
-
-        if [[ $mem_mb -gt $limit_mb ]]; then
-            log_warning "Memory limit exceeded: ${mem_mb}MB > ${limit_mb}MB"
-            return 1
-        fi
-    fi
-
-    # CPU limit check
-    if command -v ps >/dev/null 2>&1; then
-        local cpu_percent=$(ps -p "$pid" -o %cpu= 2>/dev/null | tr -d ' ' || echo "0")
-        local cpu_limit_num=$((${CPU_LIMIT%\%}))
-
-        if (( $(echo "$cpu_percent > $cpu_limit_num" | bc -l) )); then
-            log_warning "CPU limit exceeded: ${cpu_percent}% > ${CPU_LIMIT}"
-            return 1
-        fi
-    fi
-
-    return 0
-}
-
-# Start background monitoring
-start_monitoring() {
-    local pid="${1:-$$}"
-    local interval="${2:-30}"  # Check every 30 seconds
-
-    log_info "Starting process monitoring for PID $pid (interval: ${interval}s)"
-
-    # Start monitoring in background
-    (
-        while true; do
-            if ! kill -0 "$pid" 2>/dev/null; then
-                log_info "Process $pid no longer exists, stopping monitoring"
-                break
-            fi
-
-            collect_metrics "$pid"
-
-            if ! check_limits "$pid"; then
-                log_error "Resource limits exceeded, terminating process $pid"
-                kill -TERM "$pid" 2>/dev/null || true
-                break
-            fi
-
-            sleep "$interval"
-        done
-    ) &
-
-    MONITOR_PID=$!
-    echo "$MONITOR_PID"
-}
-
-# Stop monitoring
-stop_monitoring() {
-    if [[ -n "$MONITOR_PID" ]] && kill -0 "$MONITOR_PID" 2>/dev/null; then
-        kill "$MONITOR_PID" 2>/dev/null || true
-        log_info "Stopped monitoring (PID: $MONITOR_PID)"
-    fi
-}
-
-# Generate final report
-generate_report() {
-    local exit_code="${1:-0}"
-
-    if [[ -f "$METRICS_FILE" ]]; then
-        # Update with final information
-        if command -v jq >/dev/null 2>&1; then
-            jq --arg end_time "$(date -Iseconds)" \
-               --arg exit_code "$exit_code" \
-               '.end_time = $end_time | .exit_code = $exit_code' \
-               "$METRICS_FILE" > "$METRICS_FILE.tmp" && \
-               mv "$METRICS_FILE.tmp" "$METRICS_FILE"
-        fi
-
-        log_success "Process report generated: $METRICS_FILE"
-
-        # Print summary
-        if command -v jq >/dev/null 2>&1; then
-            local samples=$(jq '.samples | length' "$METRICS_FILE")
-            echo "📊 Process Metrics Summary:" >&2
-            echo "  Agent ID: $AGENT_ID" >&2
-            echo "  Samples: $samples" >&2
-            echo "  Exit Code: $exit_code" >&2
-        fi
-    fi
-}
-
-# Monitor existing process
-monitor_pid() {
-    local pid="$1"
-
-    log_info "Monitoring existing process: PID $pid"
-
-    if ! kill -0 "$pid" 2>/dev/null; then
-        log_error "Process $pid does not exist"
-        return 1
-    fi
-
-    # Start monitoring
-    local monitor_pid=$(start_monitoring "$pid")
-
-    # Wait for process to complete
-    while kill -0 "$pid" 2>/dev/null; do
-        sleep 5
-    done
-
-    # Stop monitoring
-    stop_monitoring
-
-    log_success "Process monitoring completed for PID $pid"
-}
-
-# Main execution
-main() {
-    local action="${1:-"instrument"}"
-
-    case "$action" in
-        "instrument")
-            init_telemetry
-            local monitor_pid=$(start_monitoring)
-
-            # Set up cleanup traps
-            trap 'stop_monitoring; generate_report $?' EXIT
-            trap 'stop_monitoring; generate_report 1' INT TERM
-
-            log_success "Process instrumentation started for $AGENT_ID"
-            ;;
-        "monitor-pid")
-            if [[ -z "${2:-}" ]]; then
-                log_error "PID required for monitor-pid action"
-                exit 1
-            fi
-            monitor_pid "$2"
-            ;;
-        "--help"|"-h")
-            cat << EOF
-CFN Process Instrumentation Script
-
-Usage:
-  $0                     # Instrument current process
-  $0 monitor-pid <pid>   # Monitor existing process
-  $0 --help             # Show this help
-
-Environment Variables:
-  AGENT_ID              # Agent identifier (default: hostname-PID)
-  CFN_MEMORY_LIMIT      # Memory limit (default: 2G)
-  CFN_CPU_LIMIT         # CPU limit (default: 80%)
-  CFN_TIMEOUT           # Timeout in seconds (default: 600)
-  CFN_TELEMETRY_DIR     # Telemetry storage directory
-
-This script provides process monitoring and resource limit enforcement
-for CFN Loop agents and orchestration processes.
-EOF
-            ;;
-        *)
-            log_error "Unknown action: $action"
-            exit 1
-            ;;
-    esac
-}
-
-# Execute main function if run directly
-if [[ "${BASH_SOURCE[0]}" == "${0}" ]]; then
-    main "$@"
-else
-    # When sourced, automatically instrument current process
-    init_telemetry
-    # BUG #12 FIX: Removed local to prevent bash from blocking on background process
-    # The function sets global MONITOR_PID which is used by cleanup traps
-    start_monitoring >/dev/null
-
-    # Set up cleanup traps
-    trap 'stop_monitoring; generate_report $?' EXIT
-    trap 'stop_monitoring; generate_report 1' INT TERM
-
-    log_info "Process instrumentation enabled for $AGENT_ID"
-fi