################################################################################
# SERVER TOOLKIT - CLAUDE AI CONTEXT DATABASE
################################################################################
# OPTIMIZED FOR: Claude Code AI parsing and context loading
# LAST UPDATED: 2025-10-31
# VERSION: 2.0.0
# FORMAT: Structured key-value with hierarchical sections
################################################################################

[META]
version: 2.0.0
updated: 2025-10-31
status: production_ready
base_path: /root/server-toolkit
entry_point: launcher.sh
control_panels: cPanel, Plesk, InterWorx, Standalone

[STATUS_SNAPSHOT]
# What works right now (production ready)
working:
  - modules/diagnostics/system-health-check.sh: 22-phase comprehensive diagnostics, severity scoring, baseline tracking (NEW 2025-11-01)
  - modules/security/bot-analyzer.sh: Threat scoring (0-100), CSF blocking, domain analysis
  - modules/performance/mysql-query-analyzer.sh: Slow query detection
  - lib/system-detect.sh: Auto-detect cPanel/Plesk/InterWorx, OS, web server, DB, PHP
  - lib/user-manager.sh: User selection, arrow-key menus, search, domain detection
  - lib/reference-db.sh: 1-hour cache, pipe-delimited, indexes users/domains/DBs/WordPress + health baseline
  - launcher.sh: Main menu, module execution, cleanup/reset

# What doesn't work (empty directories)
not_implemented:
  - modules/wordpress/*: 11 planned scripts
  - modules/backup/*: 7 planned scripts
  - modules/monitoring/*: 5 planned scripts
  - modules/troubleshooting/*: 9 planned scripts
  - modules/reporting/*: 6 planned scripts

[CRITICAL_RECENT_FIX]
date: 2025-10-31
bug: Domain detection showing "(no domains) (0 domains)"
root_cause: launcher.sh exported empty SYS_CONTROL_PANEL to child processes
solution: run_module() uses subshell isolation - clears all SYS_* vars before module execution
location: launcher.sh:77-84
status: verified_working
related_bugs: BUG_010, BUG_011, BUG_012, BUG_013

[ARCHITECTURE_RULES]
# DO NOT BREAK THESE - intentional design decisions
bash_strict_mode: "set -eo pipefail" (NOT -euo - -u is too strict)
grep_pattern: Always add "|| true" to grep/find that might not match
unbound_vars: Use ${var:-} or ${var:-default} for potentially unbound variables
arithmetic: Use current=$((current + 1)) NOT ((current++))
output_suppression: NEVER use { } >/dev/null on critical functions (breaks variable assignment)
module_isolation: Modules run in subshells with cleared SYS_* environment
initialization: system-detect.sh auto-runs on source, uses exec for output suppression

[KEY_FILES]
# Files you'll modify most often
launcher.sh: Main menu, runs modules via run_module() - Lines: 77-84 (subshell isolation)
lib/system-detect.sh: Auto-initialization - Lines: 433-437
lib/user-manager.sh: User selection, domain detection, search
modules/security/bot-analyzer.sh: Bot analysis with CSF blocking
config/settings.conf: User configuration
config/settings.conf.minimal: Clean template (only implemented features)
tools/test-domain-detection.sh: Quick validation of domain detection
tools/diagnostic-report.sh: System diagnostic collector
REFDB_FORMAT.txt: THIS FILE - update after changes!

[DIRECTORY_STRUCTURE]
server-toolkit/
├── launcher.sh (main entry point)
├── README.md, TROUBLESHOOTING.md, AUDIT-REPORT.md, PROJECT-STRUCTURE.md
├── config/ (settings.conf, whitelist-ips.txt, whitelist-user-agents.txt)
├── lib/ (common-functions.sh, system-detect.sh, user-manager.sh, reference-db.sh, mysql-analyzer.sh)
├── modules/
│   ├── security/ (bot-analyzer.sh ✓ WORKING)
│   ├── performance/ (mysql-query-analyzer.sh ✓ WORKING)
│   ├── wordpress/, backup/, monitoring/, troubleshooting/, reporting/ (empty - future)
└── tools/ (test-domain-detection.sh, diagnostic-report.sh)

[BUGS_FIXED]
# All bugs fixed chronologically - DO NOT REINTRODUCE

BUG_001:
  title: Launcher exits after system scan
  cause: set -euo pipefail too strict, grep/arithmetic failures
  fix: Changed to set -eo pipefail, added || true, fixed arithmetic
  files: launcher.sh:15, lib/*.sh multiple lines

BUG_002:
  title: Unbound variable errors in system-detect.sh
  cause: Checking $SYS_DETECTION_COMPLETE without default
  fix: Changed to ${SYS_DETECTION_COMPLETE:-}
  files: lib/system-detect.sh:434

BUG_003:
  title: Associative array unbound key errors
  cause: Checking seen_domains[$domain] without default
  fix: Changed to ${seen_domains[$domain]:-}
  files: lib/reference-db.sh:219,235

BUG_004:
  title: Missing functions causing build failures
  cause: get_database_owner() and get_database_domain() not defined
  fix: Added both functions
  files: lib/user-manager.sh:415-445

BUG_005:
  title: grep commands failing with set -e
  cause: grep returns exit 1 when no match
  fix: Added || true to ALL grep in pipes
  files: lib/user-manager.sh, lib/reference-db.sh

BUG_006:
  title: Arithmetic operations causing exit
  cause: ((current++)) returns exit 1 when current=0 with set -e
  fix: Changed to current=$((current + 1))
  files: lib/reference-db.sh:141,187

BUG_007:
  title: find command improper -o syntax
  cause: find -name "*.log" -o -name "*access*" needs parentheses
  fix: find \( -name "*.log" -o -name "*access*" \)
  files: lib/reference-db.sh:290 (DISABLED - performance issues)

BUG_008:
  title: Octal number error in hourly traffic timeline
  cause: Hours 08/09 treated as invalid octal
  fix: Strip leading zeros using 10#$hour
  files: modules/security/bot-analyzer.sh:1154-1157

BUG_009:
  title: User list not displaying when selecting "Specific user"
  cause: select_user_interactive() output captured by $() - all echo went to stdout
  fix: Redirect display output to stderr using { ... } >&2, only username to stdout
  files: lib/user-manager.sh:330-365,384-408

BUG_010:
  title: System detection auto-init suppressing errors
  cause: initialize_system_detection 2>/dev/null || true silently failed
  fix: Removed error suppression, let errors be visible
  files: lib/system-detect.sh:435
  lesson: Never use 2>/dev/null || true on critical functions

BUG_011:
  title: Duplicate menu display and system detection messages
  cause: Auto-init printed messages before clear command
  fix: Suppress output during auto-init with { } redirect, verbose mode via TOOLKIT_VERBOSE_INIT=1
  files: lib/system-detect.sh:433-445
  lesson: { } blocks with redirects preserve shell context

BUG_012:
  title: Cleanup/Reset not forcing fresh detection in current session
  cause: cleanup_all_data() only deleted files, not in-memory variables
  fix: Unset all SYS_* vars, re-source libraries, display new values
  files: launcher.sh:332-360

BUG_013:
  title: Brace redirection blocks variable assignment
  cause: { initialize_system_detection } >/dev/null broke variable export
  fix: Use exec file descriptor manipulation instead
  files: lib/system-detect.sh:439-445
  lesson: Brace groups with output redirect break variable assignments - use exec instead

FINAL_DOMAIN_DETECTION_FIX:
  problem: "(no domains) (0 domains)" even though pickledperil.com exists
  root_cause: launcher.sh sourced system-detect.sh, exported empty SYS_CONTROL_PANEL, child processes inherited it
  solution: run_module() runs modules in subshell that clears SYS_* before execution
  code: |
    (
        for var in $(compgen -e | grep "^SYS_"); do
            unset "$var"
        done
        "$MODULES_DIR/$category/$module" "$@"
    )
  enhancements: arrow-key menu, confirmation prompt, enhanced cleanup, test script
  verification: pickledperil - pickledperil.com (1 domains) ✓
  lessons: |
    1. Exported env vars persist across child processes
    2. Output redirect with { } can break variable assignment
    3. Subshells with ( ) provide clean environment isolation
    4. Always test exact user flow, not just isolated components

[ENHANCEMENTS]

ENHANCEMENT_001:
  title: Post-analysis menu for bot analyzer
  date: 2025-10-31
  location: modules/security/bot-analyzer.sh:1597-1662
  description: After analysis, prompts "Go to Take Action Menu or Go Back"
  behavior: |
    If high-risk IPs found (score >= 70): Offers action menu or back
    If no high-risk IPs: Success message, press Enter
  functions: show_post_analysis_menu(), calls offer_csf_blocking()

ENHANCEMENT_002:
  title: Interactive CSF blocking for high-risk IPs
  date: 2025-10-31
  location: modules/security/bot-analyzer.sh:1664-1806
  description: Auto-detects IPs with threat score >= 70, offers CSF blocking
  options: 1hr temp, 24hr temp, permanent, skip
  safety: Auto-excludes server IPs, false positives, color-coded severity
  scoring: 70-79 ELEVATED, 80-89 HIGH, 90-100 CRITICAL

ENHANCEMENT_003:
  title: Intelligent recommendation engine with domain-level analysis
  date: 2025-10-31
  location: modules/security/bot-analyzer.sh:1601-1798
  description: Context-aware recommendations based on attack scope
  analysis: Single domain, primary target, server-wide, single-server
  recommendations: IP blocking strategy, CSF CT_LIMIT, .htaccess, Apache global, WP hardening, rate limiting
  philosophy: Script suggests, user decides (user-centric, not script-centric)

ENHANCEMENT_004:
  title: Professional terminal-agnostic output design
  date: 2025-10-31
  location: lib/common-functions.sh:8-116
  changes: |
    Before: Emojis (ℹ ✓ ⚠ ✗ 🚨), fancy Unicode, 256-color
    After: Text labels [INFO] [OK] [WARNING] [ERROR] [CRITICAL], ASCII only, basic ANSI colors
  rationale: Must work everywhere - SSH, serial consoles, screen readers, piped output
  auto_detection: Disables colors if not TTY or TOOLKIT_NO_COLOR=1

ENHANCEMENT_005:
  title: User-centric action menu redesign
  date: 2025-10-31
  location: modules/security/bot-analyzer.sh:2036-2182
  changes: Shows ALL 8 actions (not just recommended), recommended marked with ⭐ and priority
  philosophy: User has full control, script is not the expert
  actions: IP blocking (1hr/24hr/perm), domain .htaccess, Apache global, CSF CT_LIMIT, rate limiting, WP hardening

ENHANCEMENT_006:
  title: User selection search functionality
  date: 2025-10-31
  location: lib/user-manager.sh:382-437
  description: Search/filter for servers with 200+ users
  usage: S [text] - case-insensitive partial match on username/domain
  behavior: |
    1 match: Confirmation prompt, auto-select on Y
    N matches: Arrow-key menu (bash select)
    0 matches: Error message
  testing: Verified with username/domain search, case-insensitive

[REFERENCE_DATABASE]
# .sysref file format - pipe-delimited database

format: TYPE|field1|field2|field3|...
location: /root/server-toolkit/.sysref
timestamp: /root/server-toolkit/.sysref.timestamp
ttl: 3600 seconds (1 hour)

record_types:
  SYS: System information
    format: SYS|key|value|extra
    example: SYS|CONTROL_PANEL|cpanel|11.130.0.15

  USER: User accounts
    format: USER|username|primary_domain|db_count|domain_count|disk_mb|home_dir
    example: USER|pickledperil|pickledperil.com|1|3|82|/home/pickledperil

  DB: Databases
    format: DB|db_name|owner|primary_domain|size_mb|table_count
    example: DB|pickledperil_wp_wt6lz|pickledperil|pickledperil.com|15.23|12

  DOMAIN: Domain mappings (ENHANCED)
    format: DOMAIN|domain|owner|doc_root|log_path|php_ver|is_primary|type|aliases
    example: DOMAIN|pickledperil.com|pickledperil|/home/pickledperil/public_html|/etc/.../pickledperil.com|ea-php81|yes|primary|www.pickledperil.com
    types: primary, addon, subdomain, alias, parked, remote

  WP: WordPress installations (ENHANCED)
    format: WP|domain|owner|path|db_name|db_user|version|plugin_count|theme_count
    example: WP|pickledperil.com|pickledperil|/home/pickledperil/public_html|pickledperil_wp_wt6lz|pickledperil_wp_user|6.8.3|1|3

  LOG: Log file locations (DISABLED - performance issues)
    format: LOG|type|domain|owner|path|size_mb
    status: Disabled, log paths now in DOMAIN records

  HEALTH: Health check baseline metrics (NEW - Added 2025-11-01)
    format: HEALTH|metric_name|value|date
    updated_by: modules/diagnostics/system-health-check.sh
    refresh: Updated every time health check runs
    examples:
      HEALTH|TIMESTAMP|2025-11-01 15:54:23|
      HEALTH|MEMORY_TOTAL_MB|3776|2025-11-01
      HEALTH|MEMORY_USED_PERCENT|45|2025-11-01
      HEALTH|CPU_LOAD_1MIN|2.4|2025-11-01
      HEALTH|CPU_CORES|2|2025-11-01
      HEALTH|DISK_USED_PERCENT|35|2025-11-01
      HEALTH|IOWAIT_PERCENT|0|2025-11-01
      HEALTH|EMAIL_QUEUE_SIZE|0|2025-11-01
      HEALTH|ZOMBIE_PROCESSES|0|2025-11-01
      HEALTH|HTTPD_STATUS|running|2025-11-01
      HEALTH|MYSQL_STATUS|running|2025-11-01
      HEALTH|FIREWALL_STATUS|csf|2025-11-01
      HEALTH|CRITICAL_ISSUES|0|2025-11-01
      HEALTH|HIGH_ISSUES|3|2025-11-01
      HEALTH|MEDIUM_ISSUES|3|2025-11-01
      HEALTH|LOW_ISSUES|1|2025-11-01
    purpose: |
      - Track system baseline metrics for trend analysis
      - Compare current vs. historical values
      - Cross-module intelligence (bot analyzer can see if memory was already high)
      - Quick reference without re-running expensive checks

query_functions:
  - db_get_user(username)
  - db_get_all_users()
  - db_get_user_databases(username)
  - db_get_user_domains(username)
  - db_get_database_owner(db_name)
  - db_get_all_wordpress()
  - db_is_fresh()
  - db_rebuild()

cross_reference_example: |
  To find everything for pickledperil.com:
  1. grep "^DOMAIN|pickledperil.com|" .sysref → get owner, doc_root, log_path, PHP version
  2. grep "^DB|.*|pickledperil|" .sysref → get all databases
  3. grep "^WP|pickledperil.com|" .sysref → get WordPress installation details
  4. grep "^DOMAIN|.*|pickledperil|" .sysref → get all subdomains/aliases

[GLOBAL_VARIABLES]

system_detect_exports:
  SYS_CONTROL_PANEL: cpanel|plesk|interworx|none
  SYS_CONTROL_PANEL_VERSION: Version string
  SYS_OS_TYPE: almalinux|centos|rhel|rocky|cloudlinux
  SYS_OS_VERSION: Version number
  SYS_WEB_SERVER: apache|nginx|litespeed|openlitespeed
  SYS_WEB_SERVER_VERSION: Version string
  SYS_DB_TYPE: mysql|mariadb|none
  SYS_DB_VERSION: Version string
  SYS_LOG_DIR: Log directory path
  SYS_USER_HOME_BASE: User home base (usually /home)
  SYS_PHP_VERSIONS: Array of installed PHP versions
  SYS_CLOUDFLARE_ACTIVE: yes|no
  SYS_CLOUDLINUX: yes|no
  CPU_CORES, CPU_USED, LOAD_AVERAGE: Resource metrics
  MEM_TOTAL, MEM_USED, MEM_PERCENT: Memory metrics

common_functions_exports:
  TOOLKIT_BASE_DIR: Base directory path
  TEMP_SESSION_DIR: Temp directory for session

reference_db_exports:
  SYSREF_DB: Path to .sysref file
  SYSREF_TIMESTAMP: Path to timestamp file

launcher_defines:
  BASE_DIR: Same as TOOLKIT_BASE_DIR
  MODULES_DIR, LIB_DIR, CONFIG_DIR: Module/library/config paths

[COMMON_PATTERNS]
# Copy-paste these for consistent code

safe_grep: |
  result=$(grep "pattern" file 2>/dev/null | head -1 || true)

safe_find: |
  files=$(find /path -name "*.txt" 2>/dev/null || true)

loop_with_progress: |
  total=${#array[@]}
  current=0
  for item in "${array[@]}"; do
      current=$((current + 1))
      show_progress $current $total "Processing..."
      # do work
  done
  finish_progress

check_module_exists: |
  if ! module_exists "category" "module.sh"; then
      print_error "Module not found"
      return 1
  fi

query_refdb: |
  db_ensure_fresh  # Rebuild if stale
  users=$(db_get_all_users)
  for user in $users; do
      domains=$(db_get_user_domains "$user")
  done

[MODULE_TEMPLATE]
# Template for creating new modules

file_location: /root/server-toolkit/modules/{category}/{name}.sh
template: |
  #!/bin/bash

  # Load common functions
  SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")/../.." && pwd)"
  source "$SCRIPT_DIR/lib/common-functions.sh"

  # Load system detection (optional)
  source "$SCRIPT_DIR/lib/system-detect.sh"

  # Load user manager (optional)
  source "$SCRIPT_DIR/lib/user-manager.sh"

  # Load reference database (optional)
  source "$SCRIPT_DIR/lib/reference-db.sh"

  print_banner "Module Name"

  # Your code here

  print_success "Completed"

steps:
  1: Create file in modules/{category}/{name}.sh
  2: chmod +x the file
  3: Add to launcher.sh handle_{category}_menu() case statement
  4: Test directly and via launcher

best_practices:
  - Use print_* functions, not echo
  - Add || true to grep/find
  - Use ${var:-default} for unbound vars
  - Use current=$((current + 1)) not ((current++))
  - Include progress indicators for long operations
  - Handle errors gracefully
  - Return proper exit codes (0=success, 1=error)

[KNOWN_ISSUES]

ISSUE_001:
  title: Log file scanning disabled
  location: lib/reference-db.sh:285
  reason: find with while read loop hangs on servers with many logs
  status: Disabled, returns empty [LOGS] section
  todo: Implement with find | head -1000 | while read

ISSUE_002:
  title: Home directory empty in USER records
  location: lib/reference-db.sh:140
  symptom: HOME_DIR field empty for some users
  cause: get_user_info may not parse cPanel files properly
  status: Low priority - doesn't break functionality
  todo: Debug get_cpanel_user_info()

ISSUE_003:
  title: No module implementation yet
  location: modules/* directories
  status: Only bot-analyzer.sh and mysql-query-analyzer.sh exist
  priority: High - need WordPress modules next
  todo: wp-cron-status.sh, wp-cron-mass-fix.sh, wp-health-check.sh

ISSUE_004:
  title: Nextcloud integration not implemented
  location: launcher.sh
  status: Config exists but download mechanism not coded
  todo: Implement module download from Nextcloud via manifest.txt

[NEXT_PRIORITIES]

immediate_high:
  1: Implement wp-cron-status.sh module
  2: Implement wp-cron-mass-fix.sh module
  3: Fix HOME_DIR empty issue in reference database
  4: Test on production server with multiple users

short_term_medium:
  5: Implement wp-health-check.sh (scan all WP installations)
  6: Implement resource-monitor.sh (performance category)
  7: Re-enable log file scanning with limits
  8: Add error handling throughout

long_term_low:
  9: Implement Nextcloud module download system
  10: Create automated daily security scan feature
  11: Build reporting dashboard
  12: Add email/Slack alert integration

[PERFORMANCE]

bottlenecks:
  - find /home -name "wp-config.php" can be slow (acceptable for cache)
  - du -sm on large home directories (acceptable for cache)
  - MySQL queries for database sizes (acceptable)
  - Log file scanning DISABLED due to excessive find operations

optimizations:
  - Reference database caching (1 hour TTL)
  - Single-pass file reading
  - Hash table lookups in bot-analyzer
  - Avoid redundant grep operations

targets:
  system_scan: < 5 seconds for typical server
  bot_analysis: < 60 seconds for 25K IPs
  menu_navigation: Instant
  module_launch: < 1 second

[TESTING_CHECKLIST]

before_commit:
  - Test with fresh system (rm .sysref*)
  - Test with existing cache
  - Test menu navigation (all categories)
  - Test module execution
  - bash -n file.sh (syntax check)
  - Check for unhandled grep failures
  - Verify exit codes
  - Test with set -e enabled
  - Verify no debug output in production

common_test_commands: |
  bash -n /root/server-toolkit/launcher.sh
  /root/server-toolkit/launcher.sh
  rm -f /root/server-toolkit/.sysref*
  /root/server-toolkit/modules/security/bot-analyzer.sh
  bash /root/server-toolkit/tools/test-domain-detection.sh

[TROUBLESHOOTING]

domain_detection_fails:
  symptom: "(no domains) (0 domains)"
  check: echo $SYS_CONTROL_PANEL (should be "cpanel")
  fix: Run cleanup option 8, or exit and restart launcher
  test: bash /root/server-toolkit/tools/test-domain-detection.sh

functions_not_found:
  symptom: "command not found" errors
  check: ls -la /root/server-toolkit/lib/*.sh
  fix: source libraries manually, check syntax with bash -n

menus_displaying_twice:
  symptom: Duplicate menus, detection messages visible
  check: TOOLKIT_VERBOSE_INIT environment variable
  fix: unset TOOLKIT_VERBOSE_INIT or set to 0

csf_not_working:
  symptom: "csf: command not found"
  check: which csf
  fix: Install CSF or use alternative security methods

[CONFIG_FILE]
# config/settings.conf

loaded_by: launcher.sh:484
format: Bash variable assignments

key_settings:
  QUICK_SCAN_HOURS: 1 (default time range for quick scans)
  AUTO_BLOCK: no (auto-apply firewall blocks)
  AUTO_BLOCK_THRESHOLD: 80 (threat score threshold)
  LOG_DIR: /var/log/apache2/domlogs (auto-detected)
  CPANEL_HOME: /home (auto-detected)
  CPU_ALERT_THRESHOLD: 80
  MEM_ALERT_THRESHOLD: 90
  DISK_ALERT_THRESHOLD: 85
  WHITELIST_IP_FILE: config/whitelist-ips.txt
  WHITELIST_UA_FILE: config/whitelist-user-agents.txt

minimal_version: config/settings.conf.minimal (only current settings, no future features)

[MENU_STRUCTURE]

main_menu:
  1: Security & Threat Analysis (bot-analyzer ✓ WORKING, 9 others ⚠️ NOT IMPLEMENTED)
  2: WordPress Management (14 modules ⚠️ ALL NOT IMPLEMENTED)
  3: Performance & Diagnostics (mysql-query-analyzer ✓ WORKING, 10 others ⚠️ NOT IMPLEMENTED)
  4: Backup & Recovery (8 modules ⚠️ ALL NOT IMPLEMENTED)
  5: Monitoring & Alerts (8 modules ⚠️ ALL NOT IMPLEMENTED)
  6: Troubleshooting (11 modules ⚠️ ALL NOT IMPLEMENTED)
  7: Reporting & Analytics (7 modules ⚠️ ALL NOT IMPLEMENTED)
  8: Cleanup/Reset ✓ WORKING
  9: Configuration ✓ WORKING
  0: Exit ✓ WORKING

security_menu_detail:
  1: Full Bot Analysis ✓ WORKING (bot-analyzer.sh)
  2: Quick Scan (1 hour) ✓ WORKING (bot-analyzer.sh -H 1)
  3-10: ⚠️ NOT IMPLEMENTED (Live Monitor, IP Lookup, Auto-Block, Whitelist, Attack Pattern, DDoS, Firewall, SSL Audit)

[FOR_NEW_CLAUDE_INSTANCES]
# Read this if you're a new Claude session

quickstart:
  1: Read [STATUS_SNAPSHOT] - know what works vs. what doesn't
  2: Read [CRITICAL_RECENT_FIX] - understand latest major fix
  3: Read [ARCHITECTURE_RULES] - don't break these intentional designs
  4: Read [BUGS_FIXED] - don't reintroduce fixed bugs
  5: Read [NEXT_PRIORITIES] - know what to work on next
  6: Test current state - run /root/server-toolkit/launcher.sh

update_after_changes:
  - Add new bugs to [BUGS_FIXED]
  - Update [STATUS_SNAPSHOT] with new working features
  - Update [NEXT_PRIORITIES]
  - Update [CRITICAL_RECENT_FIX] if applicable
  - Update [LAST UPDATED] date at top

dont_do:
  - Don't use set -u (too strict)
  - Don't suppress errors with 2>/dev/null on critical functions
  - Don't use { } blocks with output redirect (breaks variables)
  - Don't reintroduce bugs from [BUGS_FIXED]
  - Don't create new modules without following [MODULE_TEMPLATE]

[END]
# This file is the single source of truth for project context.
# Keep it updated. Keep it concise. Keep it structured.
################################################################################