From 5c63916947f76508cb990820b19630849ba110a1 Mon Sep 17 00:00:00 2001
From: Cursor Agent <cursoragent@cursor.com>
Date: Wed, 15 Jul 2026 20:24:30 +0000
Subject: [PATCH 1/8] docs: add emergency recovery guide for sliceanddice boot
 issues

Add comprehensive recovery documentation and diagnostic script for emergency
mode issues. Includes analysis of master vs development branch differences
and step-by-step recovery procedures.

Addresses potential causes:
- Missing disko.nix import from development branch
- Filesystem mounting failures
- Boot partition space issues
- Configuration branch mismatches

Co-authored-by: Alex Spaulding <aspauldingcode@gmail.com>
---
 EMERGENCY_RECOVERY.md | 130 +++++++++++++++++++++++++++++++++++
 SOLUTION.md           | 154 ++++++++++++++++++++++++++++++++++++++++++
 diagnose-emergency.sh | 108 +++++++++++++++++++++++++++++
 3 files changed, 392 insertions(+)
 create mode 100644 EMERGENCY_RECOVERY.md
 create mode 100644 SOLUTION.md
 create mode 100755 diagnose-emergency.sh

diff --git a/EMERGENCY_RECOVERY.md b/EMERGENCY_RECOVERY.md
new file mode 100644
index 00000000..c4de106c
--- /dev/null
+++ b/EMERGENCY_RECOVERY.md
@@ -0,0 +1,130 @@
+# Emergency Recovery Guide for sliceanddice
+
+## Problem Analysis
+
+Your sliceanddice NixOS system is in emergency mode. Based on the repository analysis:
+
+### Most Likely Causes:
+
+1. **Missing disko.nix import**: If you pulled/applied configuration from the `development` branch (commit `ba8fb7cb` or later), it references `./disko.nix` which doesn't exist on `master` branch.
+
+2. **Filesystem mounting failure**: The recent changes included:
+   - New disko partitioning scheme (btrfs subvolumes)
+   - Modified filesystem UUIDs and mount options
+   - New swap configuration
+
+3. **Boot generation limit**: Commit `963b83f1` capped systemd-boot generations to 5 to prevent `/boot` from filling (511M EFI partition).
+
+## Immediate Recovery Steps
+
+### Option 1: Boot into a previous generation (RECOMMENDED FIRST STEP)
+
+1. At boot, when you see the systemd-boot menu, press `Space` or `Esc` immediately
+2. Select an older NixOS generation from before your recent changes
+3. Press `Enter` to boot
+4. Once booted, you can investigate what went wrong
+
+### Option 2: Emergency mode recovery
+
+If you're already in emergency mode:
+
+```bash
+# 1. Check which filesystems are mounted
+mount | grep -E "/(boot|nix|home)"
+
+# 2. Check for filesystem errors
+journalctl -xb | grep -i "failed\|error\|emergency"
+
+# 3. Check if /boot is full
+df -h /boot
+
+# 4. Check filesystem mounts status
+systemctl list-units --type=mount --failed
+
+# 5. Try to mount root as read-write if needed
+mount -o remount,rw /
+```
+
+### Option 3: Rollback to stable master branch configuration
+
+If you're using `/etc/nixos/.dotfiles`:
+
+```bash
+# In emergency mode or from an older generation:
+cd /etc/nixos/.dotfiles
+
+# Check current branch and status
+git branch
+git status
+
+# If you're on development branch, switch to master
+git checkout master
+git pull origin master
+
+# Rebuild with stable config
+nixos-rebuild switch --flake /etc/nixos/.dotfiles#sliceanddice
+```
+
+## Known Issues from Recent Commits
+
+### Issue 1: disko.nix import on master branch
+- **Commit**: `ba8fb7cb` (on development branch only)
+- **Problem**: Imports `./disko.nix` which doesn't exist on master
+- **Solution**: Ensure you're using master branch, or cherry-pick disko.nix from development
+
+### Issue 2: Boot partition space
+- **Commit**: `963b83f1`
+- **Problem**: 511M EFI partition fills with initrds (~50MB each)
+- **Solution**: Already fixed with `boot.loader.systemd-boot.configurationLimit = 5`
+- **Check**: Run `df -h /boot` to verify space
+
+### Issue 3: Filesystem UUID changes
+- **Location**: `hosts/nixos/sliceanddice/hardware-configuration.nix`
+- **Current UUIDs**:
+  - Root: `b89f5dca-4b37-4062-bf1d-9e4ebfd61916` (ext4)
+  - Boot: `8824-4C5F` (vfat)
+  - Swap: `c570ec29-6025-456b-99d1-8f16b677835a`
+
+## Verification Commands
+
+Once you can boot normally, verify the configuration:
+
+```bash
+# Check current branch
+cd /etc/nixos/.dotfiles && git branch
+
+# Verify filesystem mounts
+findmnt / /boot /home /nix
+
+# Check boot partition space
+df -h /boot
+du -sh /boot/loader/entries/*
+
+# List available boot generations
+ls -lh /boot/loader/entries/
+
+# Check system status
+systemctl status
+journalctl -p err -b
+```
+
+## Prevention
+
+1. **Always test major changes**: Use `nixos-rebuild test` before `switch`
+2. **Keep old generations**: The configurationLimit=5 is set, but you can temporarily increase it
+3. **Monitor /boot space**: Run `df -h /boot` before rebuilds
+4. **Stick to one branch**: Use either `master` or `development`, not a mix
+
+## Contact Information
+
+If you need to restore from scratch:
+- Bootstrap script: `scripts/install-sliceanddice.sh`
+- Secrets: `secrets/sliceanddice-secrets.yaml` (encrypted with sops)
+- Hardware config: `hosts/nixos/sliceanddice/hardware-configuration.nix`
+
+## Architecture Notes
+
+- Laptop: MSI with Intel Tiger Lake UHD + NVIDIA RTX 3050 Ti Mobile
+- Storage: Samsung 870 EVO 500GB (`ata-Samsung_SSD_870_EVO_500GB_S62ANJ0R238724D`)
+- Graphics: Hybrid PRIME render offload (Intel for display, NVIDIA on-demand)
+- Kernel: Latest mainline (`pkgs.linuxPackages_latest`)
diff --git a/SOLUTION.md b/SOLUTION.md
new file mode 100644
index 00000000..88826327
--- /dev/null
+++ b/SOLUTION.md
@@ -0,0 +1,154 @@
+# Solution for sliceanddice Emergency Mode
+
+## Problem Summary
+
+Your sliceanddice NixOS system entered emergency mode. Analysis of the git repository reveals:
+
+### Root Cause
+
+The `development` branch (ahead of `master`) includes these changes to sliceanddice:
+- **NEW**: `disko.nix` - disk/filesystem configuration 
+- **NEW**: `kbd-bl-ask.nix` - keyboard backlight config
+- **NEW**: `msi-ec-sword-kbd-disable.patch` - MSI EC kernel patch
+- **MODIFIED**: `default.nix` - imports disko.nix, adds power management, nixinstall features
+- **MODIFIED**: `hardware-configuration.nix` - updated for new partition layout
+
+**If you pulled/applied configuration from `development` to your machine but the files are missing, the system can't boot.**
+
+## Immediate Recovery (Choose One)
+
+### Recovery Option A: Boot Previous Generation (FASTEST)
+
+1. **Reboot** the machine
+2. At the **systemd-boot menu**, press `Space` or `Esc`
+3. **Select an older generation** from before your changes
+4. Boot into that generation
+5. Then either:
+   - Stay on that generation (it works), OR
+   - Fix the configuration and rebuild
+
+### Recovery Option B: Fix Configuration in Emergency Mode
+
+If you're already in emergency mode shell:
+
+```bash
+# 1. Remount root as read-write
+mount -o remount,rw /
+
+# 2. Navigate to dotfiles
+cd /etc/nixos/.dotfiles
+
+# 3. Check current state
+git status
+git branch
+
+# 4. Switch to stable master branch
+git fetch origin
+git checkout master
+git reset --hard origin/master
+
+# 5. Rebuild from stable config
+nixos-rebuild switch --flake /etc/nixos/.dotfiles#sliceanddice
+
+# 6. Reboot
+reboot
+```
+
+### Recovery Option C: Use Rescue Media
+
+If the above don't work:
+
+1. Boot from NixOS install media or the nixinstall partition
+2. Mount your root filesystem:
+   ```bash
+   mount /dev/disk/by-uuid/b89f5dca-4b37-4062-bf1d-9e4ebfd61916 /mnt
+   mount /dev/disk/by-uuid/8824-4C5F /mnt/boot
+   ```
+3. Chroot and fix:
+   ```bash
+   nixos-enter --root /mnt
+   cd /etc/nixos/.dotfiles
+   git checkout master
+   nixos-rebuild switch --flake .#sliceanddice
+   exit
+   reboot
+   ```
+
+## Long-term Solution: Sync Repository
+
+To prevent this in the future, you need to decide:
+
+### Option 1: Stay on Master (Stable)
+
+Keep using the `master` branch - it's tested and working.
+
+```bash
+cd /etc/nixos/.dotfiles
+git checkout master
+git pull origin master
+nixos-rebuild switch --flake .#sliceanddice
+```
+
+### Option 2: Move to Development (New Features)
+
+If you want the new features (power management, disko, nixinstall), merge development to master:
+
+**This should be done in this repository first, then pulled to sliceanddice.**
+
+The development branch adds:
+- ✨ Disk management with disko (btrfs subvolumes)
+- ✨ Power management (RAPL/EPP, zswap, lid/suspend)
+- ✨ On-disk NixOS installer
+- ✨ Windows dual-boot support
+- ✨ Boot generation limiting (prevents /boot from filling)
+- ✨ Keyboard backlight controls
+
+## Technical Details
+
+### Current Master Branch
+- Root: ext4 (`/dev/disk/by-uuid/b89f5dca-4b37-4062-bf1d-9e4ebfd61916`)
+- No disko.nix
+- No boot generation limit
+- Basic hardware config
+
+### Development Branch Changes
+- Root: Can be btrfs with subvolumes (after reinstall via nixinstall)
+- Managed by disko.nix
+- Boot generations limited to 5
+- Power management enabled
+- Enhanced hardware support (MSI laptop features)
+
+### Why Emergency Mode Happened
+
+Most likely scenario:
+1. You ran `nixos-rebuild switch` with development branch config
+2. The config imports `./disko.nix`
+3. File doesn't exist (not pulled or wrong branch)
+4. NixOS build/activation fails
+5. System can't complete boot → emergency mode
+
+Alternative scenarios:
+- /boot partition full (unlikely with the 5-generation limit)
+- Filesystem mount failure
+- Hardware configuration mismatch
+
+## Files Provided for Recovery
+
+1. **EMERGENCY_RECOVERY.md** - Detailed recovery steps
+2. **diagnose-emergency.sh** - Diagnostic script to run in emergency mode
+3. **SOLUTION.md** - This file
+
+## Next Steps
+
+1. **First**: Boot into a previous generation (Recovery Option A)
+2. **Then**: Decide if you want to stay on master or move to development
+3. **If development**: Properly merge the branches in this repository
+4. **Finally**: Pull the correct branch to sliceanddice and rebuild
+
+## Questions?
+
+Check these files for your current configuration:
+- Hardware UUIDs: `hosts/nixos/sliceanddice/hardware-configuration.nix`
+- Main config: `hosts/nixos/sliceanddice/default.nix`
+- Boot config: In default.nix, look for `boot.loader.*`
+- Filesystems: In hardware-configuration.nix or disko.nix (if present)
diff --git a/diagnose-emergency.sh b/diagnose-emergency.sh
new file mode 100755
index 00000000..bbf04828
--- /dev/null
+++ b/diagnose-emergency.sh
@@ -0,0 +1,108 @@
+#!/usr/bin/env bash
+# Emergency diagnostic script for sliceanddice boot issues
+# Run this in emergency mode or from a rescue shell
+
+set -euo pipefail
+
+echo "=== sliceanddice Emergency Diagnostic ==="
+echo "Timestamp: $(date)"
+echo ""
+
+echo "=== System Status ==="
+systemctl status --no-pager --failed || true
+echo ""
+
+echo "=== Mount Status ==="
+mount | sort
+echo ""
+
+echo "=== Filesystem Space ==="
+df -h || true
+echo ""
+
+echo "=== Failed Mount Units ==="
+systemctl list-units --type=mount --failed --no-pager || true
+echo ""
+
+echo "=== Boot Partition Contents ==="
+if [ -d /boot ]; then
+    echo "Boot partition space:"
+    du -sh /boot/* 2>/dev/null || true
+    echo ""
+    echo "Boot entries:"
+    ls -lh /boot/loader/entries/ 2>/dev/null || true
+else
+    echo "/boot not mounted"
+fi
+echo ""
+
+echo "=== Filesystem UUIDs ==="
+echo "Expected UUIDs from hardware-configuration.nix:"
+echo "  Root (ext4): b89f5dca-4b37-4062-bf1d-9e4ebfd61916"
+echo "  Boot (vfat): 8824-4C5F"
+echo "  Swap:        c570ec29-6025-456b-99d1-8f16b677835a"
+echo ""
+echo "Actual block devices:"
+lsblk -o NAME,UUID,FSTYPE,SIZE,MOUNTPOINT || true
+echo ""
+
+echo "=== Git Configuration Status ==="
+if [ -d /etc/nixos/.dotfiles ]; then
+    cd /etc/nixos/.dotfiles
+    echo "Branch: $(git branch --show-current 2>/dev/null || echo 'unknown')"
+    echo "Last commit: $(git log -1 --oneline 2>/dev/null || echo 'unknown')"
+    echo "Status:"
+    git status -s 2>/dev/null || true
+    echo ""
+    echo "Checking for disko.nix:"
+    if [ -f hosts/nixos/sliceanddice/disko.nix ]; then
+        echo "  ✓ disko.nix exists"
+    else
+        echo "  ✗ disko.nix MISSING"
+        echo "  → This is likely the problem if default.nix imports it!"
+    fi
+    echo ""
+    echo "Checking imports in default.nix:"
+    grep -E "^\s*(./disko\.nix|imports\s*=)" hosts/nixos/sliceanddice/default.nix 2>/dev/null || true
+else
+    echo "/etc/nixos/.dotfiles not found"
+fi
+echo ""
+
+echo "=== Recent Boot Logs (errors only) ==="
+journalctl -p err -b --no-pager -n 50 2>/dev/null || true
+echo ""
+
+echo "=== Kernel Command Line ==="
+cat /proc/cmdline 2>/dev/null || true
+echo ""
+
+echo "=== Available Generations ==="
+if [ -d /nix/var/nix/profiles ]; then
+    ls -lh /nix/var/nix/profiles/system-*-link 2>/dev/null | tail -10 || true
+else
+    echo "Profile directory not accessible"
+fi
+echo ""
+
+echo "=== Suggested Actions ==="
+echo ""
+echo "1. If disko.nix is missing but imported:"
+echo "   cd /etc/nixos/.dotfiles && git checkout master"
+echo ""
+echo "2. If /boot is full:"
+echo "   nix-collect-garbage -d"
+echo "   Or temporarily increase configurationLimit"
+echo ""
+echo "3. If filesystems won't mount:"
+echo "   Check 'journalctl -xb' for specific mount errors"
+echo "   Verify UUIDs match: blkid /dev/sdX"
+echo ""
+echo "4. Rollback to previous generation:"
+echo "   From systemd-boot menu, select an older entry"
+echo ""
+echo "5. Force rebuild from master:"
+echo "   cd /etc/nixos/.dotfiles"
+echo "   git checkout master && git reset --hard origin/master"
+echo "   nixos-rebuild switch --flake /etc/nixos/.dotfiles#sliceanddice"
+echo ""

From 568b8f8104c24c978a291fbf0e867bce604f2210 Mon Sep 17 00:00:00 2001
From: Cursor Agent <cursoragent@cursor.com>
Date: Wed, 15 Jul 2026 20:24:52 +0000
Subject: [PATCH 2/8] docs: add quick-start recovery guide

Co-authored-by: Alex Spaulding <aspauldingcode@gmail.com>
---
 QUICKSTART.md | 94 +++++++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 94 insertions(+)
 create mode 100644 QUICKSTART.md

diff --git a/QUICKSTART.md b/QUICKSTART.md
new file mode 100644
index 00000000..c0d13c8f
--- /dev/null
+++ b/QUICKSTART.md
@@ -0,0 +1,94 @@
+# Quick Recovery Guide - sliceanddice Emergency Mode
+
+## Fastest Solution (Do This First!)
+
+### Step 1: Boot Previous Generation
+
+1. **Reboot** your sliceanddice machine
+2. When you see the boot menu, press **Space** or **Esc** 
+3. **Select an older NixOS generation** (one from before the issue)
+4. Press Enter to boot
+
+That's it! You should now be back in a working system.
+
+---
+
+## Step 2: Prevent It From Happening Again
+
+Once you're booted into the working generation:
+
+```bash
+# Go to your dotfiles
+cd /etc/nixos/.dotfiles
+
+# Check which branch you're on
+git branch
+# If you see "development", that's likely the issue
+
+# Switch to stable master branch
+git checkout master
+git pull origin master
+
+# Make this the current system
+sudo nixos-rebuild switch --flake /etc/nixos/.dotfiles#sliceanddice
+
+# Reboot to test
+sudo reboot
+```
+
+---
+
+## What Happened?
+
+The `development` branch has new features but requires additional files:
+- `disko.nix` - disk configuration
+- `kbd-bl-ask.nix` - keyboard backlight
+- MSI-specific kernel patches
+- Power management features
+
+If you rebuilt with the development branch but those files weren't present, NixOS couldn't complete the build and entered emergency mode.
+
+---
+
+## Alternative: Fix in Emergency Mode
+
+If you can't access the boot menu:
+
+```bash
+# You should be at an emergency shell prompt
+# Remount root as writable
+mount -o remount,rw /
+
+# Fix the configuration
+cd /etc/nixos/.dotfiles
+git checkout master
+git reset --hard origin/master
+
+# Rebuild
+nixos-rebuild switch --flake .#sliceanddice
+
+# Reboot
+reboot
+```
+
+---
+
+## Need More Help?
+
+See the detailed guides:
+- **SOLUTION.md** - Full explanation and multiple recovery options
+- **EMERGENCY_RECOVERY.md** - Detailed technical recovery procedures  
+- **diagnose-emergency.sh** - Run this script to diagnose the issue
+
+Or run the diagnostic script:
+```bash
+bash diagnose-emergency.sh
+```
+
+---
+
+## Summary
+
+**Problem**: Configuration mismatch between branches  
+**Solution**: Boot previous generation, switch to master branch  
+**Prevention**: Stick to one branch (master is stable)

From 4da133ecf15a72c8e456f8ae2af6d6c22069bbd4 Mon Sep 17 00:00:00 2001
From: Cursor Agent <cursoragent@cursor.com>
Date: Wed, 15 Jul 2026 20:25:49 +0000
Subject: [PATCH 3/8] feat: add automatic emergency recovery script

Provides one-command fix for emergency mode issues. The script:
- Detects configuration problems (missing disko.nix, wrong branch)
- Switches to stable master branch
- Rebuilds the system configuration
- Provides clear feedback and next steps

Usage: sudo bash fix-emergency-auto.sh

Co-authored-by: Alex Spaulding <aspauldingcode@gmail.com>
---
 fix-emergency-auto.sh | 167 ++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 167 insertions(+)
 create mode 100755 fix-emergency-auto.sh

diff --git a/fix-emergency-auto.sh b/fix-emergency-auto.sh
new file mode 100755
index 00000000..5fd9132e
--- /dev/null
+++ b/fix-emergency-auto.sh
@@ -0,0 +1,167 @@
+#!/usr/bin/env bash
+# Automatic emergency recovery for sliceanddice
+# Run this script to automatically fix the emergency mode issue
+#
+# Usage:
+#   bash fix-emergency-auto.sh [--force]
+
+set -euo pipefail
+
+FORCE=false
+if [[ "${1:-}" == "--force" ]]; then
+    FORCE=true
+fi
+
+DOTFILES_DIR="/etc/nixos/.dotfiles"
+HOST="sliceanddice"
+
+echo "╔════════════════════════════════════════════════════════════════╗"
+echo "║     sliceanddice Emergency Recovery - Automatic Fixer         ║"
+echo "╚════════════════════════════════════════════════════════════════╝"
+echo ""
+
+# Check if we're running as root
+if [[ $EUID -ne 0 ]]; then
+    echo "❌ This script must be run as root (use sudo)"
+    exit 1
+fi
+
+# Check if dotfiles directory exists
+if [[ ! -d "$DOTFILES_DIR" ]]; then
+    echo "❌ Dotfiles directory not found: $DOTFILES_DIR"
+    echo "   Expected location: /etc/nixos/.dotfiles"
+    exit 1
+fi
+
+cd "$DOTFILES_DIR"
+
+echo "📍 Current location: $(pwd)"
+echo ""
+
+# Check git status
+echo "🔍 Checking git status..."
+CURRENT_BRANCH=$(git branch --show-current 2>/dev/null || echo "unknown")
+echo "   Current branch: $CURRENT_BRANCH"
+
+# Check if disko.nix exists
+DISKO_EXISTS=false
+if [[ -f "hosts/nixos/sliceanddice/disko.nix" ]]; then
+    DISKO_EXISTS=true
+    echo "   ✓ disko.nix exists"
+else
+    echo "   ✗ disko.nix missing"
+fi
+
+# Check if default.nix imports disko
+IMPORTS_DISKO=false
+if grep -q "./disko.nix" hosts/nixos/sliceanddice/default.nix 2>/dev/null; then
+    IMPORTS_DISKO=true
+    echo "   ⚠ default.nix imports disko.nix"
+else
+    echo "   ✓ default.nix does NOT import disko.nix"
+fi
+
+echo ""
+
+# Determine if we need to fix
+NEED_FIX=false
+FIX_REASON=""
+
+if [[ "$IMPORTS_DISKO" == true && "$DISKO_EXISTS" == false ]]; then
+    NEED_FIX=true
+    FIX_REASON="default.nix imports disko.nix but file is missing"
+elif [[ "$CURRENT_BRANCH" == "development" ]]; then
+    NEED_FIX=true
+    FIX_REASON="on development branch which may have incomplete features"
+fi
+
+if [[ "$NEED_FIX" == false ]]; then
+    echo "✅ Configuration appears correct!"
+    echo "   No automatic fix needed."
+    echo ""
+    echo "   If you're still experiencing issues, see:"
+    echo "   - EMERGENCY_RECOVERY.md for manual steps"
+    echo "   - diagnose-emergency.sh for diagnostics"
+    exit 0
+fi
+
+echo "⚠️  Issue detected: $FIX_REASON"
+echo ""
+
+if [[ "$FORCE" == false ]]; then
+    echo "Proposed fix:"
+    echo "  1. Fetch latest from origin"
+    echo "  2. Switch to stable master branch"
+    echo "  3. Reset to origin/master"
+    echo "  4. Rebuild NixOS configuration"
+    echo ""
+    read -p "Apply this fix? (y/N) " -n 1 -r
+    echo ""
+    if [[ ! $REPLY =~ ^[Yy]$ ]]; then
+        echo "❌ Aborted by user"
+        exit 1
+    fi
+fi
+
+echo ""
+echo "🔧 Applying fix..."
+echo ""
+
+# Step 1: Fetch
+echo "→ Fetching from origin..."
+if ! git fetch origin; then
+    echo "⚠️  Git fetch failed. Continuing anyway..."
+fi
+
+# Step 2: Checkout master
+echo "→ Switching to master branch..."
+if ! git checkout master; then
+    echo "❌ Failed to checkout master branch"
+    exit 1
+fi
+
+# Step 3: Reset to origin/master
+echo "→ Resetting to origin/master..."
+if ! git reset --hard origin/master; then
+    echo "⚠️  Git reset failed. Continuing anyway..."
+fi
+
+# Step 4: Show current state
+echo ""
+echo "📊 Current configuration:"
+echo "   Branch: $(git branch --show-current)"
+echo "   Commit: $(git log -1 --oneline)"
+echo ""
+
+# Step 5: Rebuild
+echo "→ Rebuilding NixOS configuration..."
+echo "   This may take several minutes..."
+echo ""
+
+if nixos-rebuild switch --flake "$DOTFILES_DIR#$HOST"; then
+    echo ""
+    echo "╔════════════════════════════════════════════════════════════════╗"
+    echo "║                  ✅ FIX APPLIED SUCCESSFULLY!                  ║"
+    echo "╚════════════════════════════════════════════════════════════════╝"
+    echo ""
+    echo "Next steps:"
+    echo "  1. Verify the system is working correctly"
+    echo "  2. Reboot to ensure it boots cleanly: sudo reboot"
+    echo ""
+    echo "Your system should now be running on the stable master branch."
+else
+    echo ""
+    echo "❌ nixos-rebuild failed!"
+    echo ""
+    echo "This could mean:"
+    echo "  - Network issues preventing package downloads"
+    echo "  - Evaluation errors in the configuration"
+    echo "  - Hardware issues"
+    echo ""
+    echo "Next steps:"
+    echo "  1. Check the error messages above"
+    echo "  2. Try rebooting to a previous generation"
+    echo "  3. See EMERGENCY_RECOVERY.md for manual recovery"
+    echo ""
+    exit 1
+fi

From 303452a22d8a3f0d62da4eba70dff28bd67c9c2c Mon Sep 17 00:00:00 2001
From: Cursor Agent <cursoragent@cursor.com>
Date: Wed, 15 Jul 2026 20:26:32 +0000
Subject: [PATCH 4/8] docs: add main emergency recovery README

Central emergency recovery documentation that ties all the recovery
resources together. Provides quick navigation to the right solution
based on user's situation and technical comfort level.

Co-authored-by: Alex Spaulding <aspauldingcode@gmail.com>
---
 README-EMERGENCY.md | 173 ++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 173 insertions(+)
 create mode 100644 README-EMERGENCY.md

diff --git a/README-EMERGENCY.md b/README-EMERGENCY.md
new file mode 100644
index 00000000..dd804bd2
--- /dev/null
+++ b/README-EMERGENCY.md
@@ -0,0 +1,173 @@
+# 🆘 sliceanddice Emergency Mode - Recovery Instructions
+
+> **STATUS**: Your sliceanddice NixOS system is in emergency mode  
+> **CAUSE**: Configuration mismatch between master and development branches  
+> **SOLUTION**: Multiple recovery paths available (choose the easiest one for you)
+
+---
+
+## 🚀 Quick Fix (Recommended - 2 minutes)
+
+### Option 1: Boot Previous Generation (FASTEST!)
+
+1. **Reboot** the sliceanddice machine
+2. At the systemd-boot menu, press **`Space`** or **`Esc`**
+3. Select an **older NixOS generation** (from before the problem)
+4. Press **Enter** to boot
+
+✅ **Done!** You're back in a working system.
+
+Then prevent it from happening again:
+```bash
+cd /etc/nixos/.dotfiles
+git checkout master
+sudo nixos-rebuild switch --flake .#sliceanddice
+```
+
+---
+
+## 🔧 Automated Fix Script
+
+If you're in a working shell (emergency mode or booted from old generation):
+
+```bash
+cd /etc/nixos/.dotfiles
+sudo bash fix-emergency-auto.sh
+```
+
+This script will:
+- ✅ Detect the configuration problem
+- ✅ Switch to stable master branch  
+- ✅ Rebuild your system
+- ✅ Verify everything works
+
+---
+
+## 📖 Detailed Documentation
+
+Choose the guide that fits your situation:
+
+| File | Purpose | When to Use |
+|------|---------|-------------|
+| **QUICKSTART.md** | Fast recovery steps | You want the TL;DR |
+| **SOLUTION.md** | Complete explanation | You want to understand what happened |
+| **EMERGENCY_RECOVERY.md** | Technical deep-dive | You need detailed recovery procedures |
+| **diagnose-emergency.sh** | Diagnostic script | You want to see what's wrong |
+| **fix-emergency-auto.sh** | Automated fix | You want one-command recovery |
+
+---
+
+## 🔍 What Happened?
+
+**Short version**: The `development` branch has new features that require additional files. If you rebuilt using development branch config but the files weren't present, NixOS couldn't activate and dropped into emergency mode.
+
+**Files required by development branch**:
+- `disko.nix` - Disk/filesystem management
+- `kbd-bl-ask.nix` - Keyboard backlight config
+- `msi-ec-sword-kbd-disable.patch` - MSI laptop kernel patch
+- Power management features
+- Boot generation limiting
+
+**Current master branch**: ✅ Stable, working, doesn't need these files
+
+---
+
+## 🛠️ Recovery Paths
+
+### Path A: Previous Generation (Easiest)
+Boot from systemd-boot menu → Select older generation → Boot
+
+### Path B: Emergency Shell Fix
+```bash
+mount -o remount,rw /
+cd /etc/nixos/.dotfiles
+git checkout master
+nixos-rebuild switch --flake .#sliceanddice
+reboot
+```
+
+### Path C: Automated Script
+```bash
+sudo bash fix-emergency-auto.sh
+```
+
+### Path D: Rescue Media
+Boot NixOS installer → Mount filesystems → Chroot → Fix config
+
+---
+
+## ⚡ After Recovery
+
+Once your system is working again:
+
+1. **Verify you're on master branch**:
+   ```bash
+   cd /etc/nixos/.dotfiles
+   git branch  # Should show: * master
+   ```
+
+2. **Stay on master** (recommended):
+   - It's stable and tested
+   - No configuration changes needed
+   - Everything works
+
+3. **OR move to development** (advanced):
+   - Merge development features properly
+   - Test in VM first
+   - Ensure all files are present
+
+---
+
+## 📞 Still Having Issues?
+
+1. Run the diagnostic script:
+   ```bash
+   bash diagnose-emergency.sh
+   ```
+
+2. Check the detailed guides:
+   - Read `SOLUTION.md` for full explanation
+   - Read `EMERGENCY_RECOVERY.md` for technical details
+
+3. Verify filesystem integrity:
+   ```bash
+   lsblk -o NAME,UUID,FSTYPE,SIZE,MOUNTPOINT
+   df -h
+   mount | grep -E "/(boot|nix|home)"
+   ```
+
+---
+
+## 📊 System Information
+
+**Machine**: MSI laptop with Intel Tiger Lake + NVIDIA RTX 3050 Ti  
+**Storage**: Samsung 870 EVO 500GB  
+**Current filesystem**: ext4 on root  
+**Boot**: systemd-boot (511M EFI partition)  
+
+**Expected UUIDs**:
+- Root: `b89f5dca-4b37-4062-bf1d-9e4ebfd61916`
+- Boot: `8824-4C5F`
+- Swap: `c570ec29-6025-456b-99d1-8f16b677835a`
+
+---
+
+## ✅ Success Indicators
+
+You'll know recovery worked when:
+- ✅ System boots to desktop without emergency mode
+- ✅ All applications work normally
+- ✅ `systemctl status` shows no critical failures
+- ✅ You're on master branch: `git branch` in dotfiles shows `* master`
+
+---
+
+## 🎯 Summary
+
+| Problem | Fix | Time |
+|---------|-----|------|
+| Emergency mode | Boot previous generation | 1 min |
+| + Want to prevent | Switch to master branch | 5 min |
+| + Want automated fix | Run fix-emergency-auto.sh | 10 min |
+
+**Bottom line**: Boot an older generation from the boot menu. Problem solved! 🎉

From 347940f7a2f5715cce7b3fd53017b274229d5bd4 Mon Sep 17 00:00:00 2001
From: Cursor Agent <cursoragent@cursor.com>
Date: Wed, 15 Jul 2026 20:26:57 +0000
Subject: [PATCH 5/8] docs: add emergency recovery link to main README

Co-authored-by: Alex Spaulding <aspauldingcode@gmail.com>
---
 README.md | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/README.md b/README.md
index 5cee881a..8a1a9a0c 100644
--- a/README.md
+++ b/README.md
@@ -2,6 +2,8 @@
 
 Personal, declarative system configuration built by **Alex Spaulding (aspauldingcode)** using Nix Flakes.
 
+> **🆘 EMERGENCY RECOVERY**: If your sliceanddice system is in emergency mode, see [`README-EMERGENCY.md`](README-EMERGENCY.md) for immediate recovery instructions.
+
 <!-- fleet-status:start -->
 
 ## Fleet

From ee5120818de6abd6c0f33e72332ece3b8d8a49f8 Mon Sep 17 00:00:00 2001
From: Cursor Agent <cursoragent@cursor.com>
Date: Wed, 15 Jul 2026 20:27:43 +0000
Subject: [PATCH 6/8] docs: add START_HERE quick guide for immediate action

Co-authored-by: Alex Spaulding <aspauldingcode@gmail.com>
---
 START_HERE.md | 201 ++++++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 201 insertions(+)
 create mode 100644 START_HERE.md

diff --git a/START_HERE.md b/START_HERE.md
new file mode 100644
index 00000000..5b46364b
--- /dev/null
+++ b/START_HERE.md
@@ -0,0 +1,201 @@
+# 🚨 START HERE - sliceanddice Emergency Recovery
+
+**YOUR SYSTEM**: sliceanddice (NixOS laptop)  
+**STATUS**: Emergency mode  
+**URGENCY**: Medium (system is recoverable, data is safe)  
+**SOLUTION TIME**: 1-10 minutes depending on method
+
+---
+
+## What You Need to Do RIGHT NOW
+
+### The 1-Minute Fix (DO THIS FIRST!)
+
+1. **Reboot** your sliceanddice laptop
+2. You'll see a boot menu with several options
+3. Press **`Space`** or **`Esc`** key
+4. Use arrow keys to select an **older entry** (one from before the issue)
+5. Press **`Enter`** to boot
+
+**That's it!** Your system should boot normally now.
+
+---
+
+## Understanding What Happened
+
+### The Problem
+Your NixOS configuration repository has two branches:
+- **`master`** - Stable, working configuration ✅
+- **`development`** - New features, requires extra files ⚠️
+
+You recently made changes that may have mixed these branches, causing the system to look for files that don't exist, which triggered emergency mode.
+
+### The Branches
+
+**Master Branch (current):**
+- Simple, stable configuration
+- ext4 filesystem
+- Works out of the box
+
+**Development Branch (ahead):**
+- Adds `disko.nix` for disk management
+- Adds `kbd-bl-ask.nix` for keyboard backlight
+- Adds power management features
+- Adds on-disk installer
+- **Requires** all these files to exist
+
+---
+
+## After You've Booted Successfully
+
+Once you're back in a working system, **prevent it from happening again**:
+
+```bash
+# Open a terminal
+cd /etc/nixos/.dotfiles
+
+# Check which branch you're on
+git branch
+# If it shows: * development (or anything other than master)
+
+# Switch to stable master
+git checkout master
+git pull origin master
+
+# Make this the active configuration
+sudo nixos-rebuild switch --flake /etc/nixos/.dotfiles#sliceanddice
+
+# Verify it worked
+sudo reboot
+```
+
+---
+
+## Alternative: Use the Automatic Fix Script
+
+If you're comfortable with automation:
+
+```bash
+cd /etc/nixos/.dotfiles
+sudo bash fix-emergency-auto.sh
+```
+
+This script will:
+1. Detect the problem
+2. Switch to master branch
+3. Rebuild your system
+4. Tell you if it worked
+
+---
+
+## All Available Resources
+
+I've created comprehensive documentation for you:
+
+| File | Purpose | Read When |
+|------|---------|-----------|
+| **START_HERE.md** | Quick start (this file) | Right now! |
+| **README-EMERGENCY.md** | Main recovery hub | After quick fix |
+| **QUICKSTART.md** | Fast recovery steps | You want TL;DR |
+| **SOLUTION.md** | Full explanation | You want details |
+| **EMERGENCY_RECOVERY.md** | Technical deep-dive | You need advanced help |
+| **diagnose-emergency.sh** | Diagnostic tool | Something's still wrong |
+| **fix-emergency-auto.sh** | Automated fix | You want one command |
+
+---
+
+## Your Recovery Path
+
+```
+┌─────────────────────────┐
+│ You are here:           │
+│ System in emergency mode│
+└───────────┬─────────────┘
+            │
+            ├── Option 1: Boot old generation (1 min) ──┐
+            │                                            │
+            ├── Option 2: Auto-fix script (5 min) ──────┤
+            │                                            │
+            └── Option 3: Manual fix (10 min) ──────────┤
+                                                         │
+                                            ┌────────────▼───────────┐
+                                            │ System working again   │
+                                            │ On stable master branch│
+                                            └────────────────────────┘
+```
+
+---
+
+## Technical Summary (For Reference)
+
+### Your System Specs
+- **Machine**: MSI laptop (Intel Tiger Lake + NVIDIA RTX 3050 Ti)
+- **Storage**: Samsung 870 EVO 500GB
+- **Current FS**: ext4
+- **Boot**: systemd-boot (511M EFI)
+
+### Root Cause Analysis
+- Development branch refs `./disko.nix` import
+- File exists in development but not in master
+- If system tried to build with development config on master branch
+- → Missing file → Build/activation failure → Emergency mode
+
+### Expected Filesystem UUIDs
+- Root: `b89f5dca-4b37-4062-bf1d-9e4ebfd61916` (ext4)
+- Boot: `8824-4C5F` (vfat)
+- Swap: `c570ec29-6025-456b-99d1-8f16b677835a`
+
+---
+
+## FAQ
+
+**Q: Will I lose data?**  
+A: No! Your data is safe. This is a configuration issue, not a data loss issue.
+
+**Q: Can I just stay on the old generation?**  
+A: Yes! It's perfectly fine to boot the old generation and use that. You don't have to "fix" anything if the old generation works for you.
+
+**Q: Do I have to switch to master branch?**  
+A: For now, yes, unless you want to properly merge the development features (which requires more work).
+
+**Q: What if the old generation doesn't work?**  
+A: See `EMERGENCY_RECOVERY.md` for rescue media instructions. But this is very unlikely.
+
+**Q: Can I keep using development branch?**  
+A: Yes, but you need to ensure ALL the required files are present and properly tracked in git.
+
+---
+
+## Next Steps
+
+1. ✅ **Now**: Boot previous generation (the 1-minute fix above)
+2. ✅ **Next**: Switch to master branch (commands above)
+3. ✅ **Later**: Read `README-EMERGENCY.md` to understand everything
+4. ✅ **Future**: Decide if you want to stay on master or properly setup development
+
+---
+
+## Pull Request
+
+All this documentation has been added to your repository in this PR:
+- **Branch**: `cursor/fix-sliceanddice-emergency-mode-75c8`
+- **PR**: #188 on GitHub
+
+You can merge it once your system is recovered to have these docs available for future reference.
+
+---
+
+## Support
+
+If you're still stuck after trying the above:
+
+1. Run diagnostic: `bash diagnose-emergency.sh`
+2. Read detailed guides: `EMERGENCY_RECOVERY.md`, `SOLUTION.md`
+3. Check system logs: `journalctl -xb | grep -i error`
+4. Verify mounts: `mount | grep -E "/(boot|nix)"`
+
+**Remember**: Your data is safe. This is fixable. The fastest path is: Reboot → Boot old generation → Done!
+
+---
+
+Good luck! 🍀

From 2c98be7bd4ff86f67a84a59e7ca6b24879cf119b Mon Sep 17 00:00:00 2001
From: Cursor Agent <cursoragent@cursor.com>
Date: Wed, 11 Nov 2026 17:58:29 +0000
Subject: [PATCH 7/8] feat: add TTY1 greeter and Cursor multi-instance
 diagnostics

Add comprehensive debugging and fix guides for:
- TTY1 blank cursor / gtkgreet not starting
- Cursor running multiple instances
- Wayland/XWayland display issues

Includes diagnostic scripts and step-by-step troubleshooting.

Co-authored-by: Alex Spaulding <aspauldingcode@gmail.com>
---
 FIX-CURSOR-MULTIPLE-INSTANCES.md | 456 +++++++++++++++++++++++++++++++
 FIX-TTY1-GREETER.md              | 296 ++++++++++++++++++++
 diagnose-tty1-greeter.sh         | 157 +++++++++++
 3 files changed, 909 insertions(+)
 create mode 100644 FIX-CURSOR-MULTIPLE-INSTANCES.md
 create mode 100644 FIX-TTY1-GREETER.md
 create mode 100755 diagnose-tty1-greeter.sh

diff --git a/FIX-CURSOR-MULTIPLE-INSTANCES.md b/FIX-CURSOR-MULTIPLE-INSTANCES.md
new file mode 100644
index 00000000..95dde140
--- /dev/null
+++ b/FIX-CURSOR-MULTIPLE-INSTANCES.md
@@ -0,0 +1,456 @@
+# Fix: Cursor Multiple Instances / Display Issues
+
+## Problem Summary
+
+Cursor is running multiple instances and not displaying correctly. An isolated launch works, but the normal instance doesn't show windows properly.
+
+## Root Cause Analysis
+
+Based on diagnostics, the issue is **NOT** a broken Cursor installation or NixOS configuration. Instead:
+
+### Confirmed Issues:
+1. **Existing Cursor process (PID 23387)** - New launches forward to already-running instance
+2. **Isolated profile works** - Fresh user-data directory launches successfully  
+3. **Existing renderers use X11** - Original instance running through XWayland
+4. **Mesa `Permission denied`** - GBM driver load failure (impact unknown)
+5. **Stuck window state** - Process running but not displaying windows
+
+### Why Isolated Launch Worked
+
+The successful isolated launch command changed three things:
+1. **Fresh user-data directory** - Bypassed `~/.config/Cursor`
+2. **Disabled extensions** - Prevented extensions from loading
+3. **Native Wayland** - Requested Wayland instead of X11 backend
+
+**Conclusion**: The problem is in your **existing Cursor profile/state**, not the application itself.
+
+## Investigation Summary Table
+
+| Finding                             | Meaning                                                                         | Assessment                      |
+| ----------------------------------- | ------------------------------------------------------------------------------- | ------------------------------- |
+| Existing Cursor process (PID 23387) | New launches were forwarded to an already-running instance                      | ✅ Confirmed                    |
+| Isolated profile works              | Cursor can launch with a fresh user-data directory                              | ✅ Confirmed                    |
+| Existing renderers use X11          | Your original instance was running through XWayland                             | ✅ Confirmed                    |
+| Mesa `Permission denied`            | A GBM driver could not be loaded                                                | ⚠️ Confirmed error, impact unknown |
+| Four Electron flag warnings         | Cursor doesn't recognize those options in its application-level argument parser | ℹ️ Not necessarily fatal        |
+| `XDG_SESSION_TYPE=tty`              | Your graphical session may have been launched from a TTY                        | ℹ️ Not necessarily an issue     |
+
+## Step-by-Step Fix
+
+### Step 1: Kill All Cursor Instances
+
+```bash
+# Find all Cursor processes
+ps aux | grep -i cursor
+
+# Kill them all
+pkill -9 cursor
+pkill -9 Cursor
+
+# Verify they're gone
+ps aux | grep -i cursor
+```
+
+### Step 2: Test Clean Launch
+
+After killing all instances, test a normal launch with Wayland:
+
+```bash
+cursor --ozone-platform=wayland
+```
+
+**If this works**: The problem was the stuck process. You're done!
+
+**If this fails**: Continue to Step 3.
+
+### Step 3: Test Without Extensions
+
+Close Cursor and try launching without extensions:
+
+```bash
+cursor --disable-extensions --ozone-platform=wayland
+```
+
+**If this works**: An extension is causing the issue. See [Extension Debugging](#extension-debugging) below.
+
+**If this fails**: Continue to Step 4.
+
+### Step 4: Test GPU Configuration
+
+Close Cursor and try with disabled GPU acceleration:
+
+```bash
+cursor --disable-gpu --ozone-platform=wayland
+```
+
+**If this works**: GPU/Mesa configuration issue. See [GPU Debugging](#gpu-debugging) below.
+
+**If this fails**: Continue to Step 5.
+
+### Step 5: Test Fresh Profile (Persistent)
+
+Create a test profile and use it:
+
+```bash
+mkdir -p ~/.local/share/cursor-test-profile
+cursor --user-data-dir="$HOME/.local/share/cursor-test-profile" --ozone-platform=wayland
+```
+
+**If this works**: Your original profile data is corrupted. See [Profile Repair](#profile-repair) below.
+
+**If this fails**: The issue is deeper. See [Advanced Debugging](#advanced-debugging).
+
+## Extension Debugging
+
+If Step 3 worked (no extensions = works), identify the problematic extension:
+
+```bash
+# 1. Launch Cursor without extensions
+cursor --disable-extensions --ozone-platform=wayland
+
+# 2. Open Extensions view (Ctrl+Shift+X)
+
+# 3. Disable all extensions manually
+
+# 4. Close Cursor and relaunch normally
+cursor --ozone-platform=wayland
+
+# 5. Enable extensions one-by-one to find the culprit
+```
+
+**Common problematic extensions**:
+- Remote development extensions
+- Window management extensions
+- Custom UI/theme extensions
+- Extensions that hook into window creation
+
+## GPU Debugging
+
+If Step 4 worked (disabled GPU = works), investigate graphics issues:
+
+### Check Mesa/GBM Setup
+
+```bash
+# Check for Mesa errors
+journalctl --user -u cursor -n 100 | grep -i mesa
+
+# Verify DRI devices
+ls -la /dev/dri/
+
+# Check OpenGL
+glxinfo | grep -i "opengl\|vendor"
+```
+
+### Check NVIDIA/Intel Hybrid Graphics
+
+Your sliceanddice has **Intel + NVIDIA hybrid graphics**. Cursor might be trying to use the wrong GPU.
+
+```bash
+# Force Intel GPU (integrated)
+__GLX_VENDOR_LIBRARY_NAME=mesa cursor --ozone-platform=wayland
+
+# Or force NVIDIA GPU (requires nvidia-offload)
+nvidia-offload cursor --ozone-platform=wayland
+```
+
+### Wayland vs XWayland
+
+The logs show your original instance used X11 through XWayland. Try forcing native Wayland:
+
+```bash
+# Ensure Wayland flags are set
+export NIXOS_OZONE_WL=1
+cursor --enable-features=UseOzonePlatform --ozone-platform=wayland
+```
+
+## Profile Repair
+
+If Step 5 worked (fresh profile = works), your original profile has issues. **Don't delete everything!** Instead, selectively repair:
+
+### Option 1: Reset Window State (Safest)
+
+```bash
+# Backup first
+cp -r ~/.config/Cursor ~/.config/Cursor.backup
+
+# Remove window state files
+rm -f ~/.config/Cursor/User/globalStorage/state.vscdb*
+rm -f ~/.config/Cursor/User/workspaceStorage/*/state.vscdb*
+
+# Try launching
+cursor --ozone-platform=wayland
+```
+
+### Option 2: Disable Cached Data
+
+```bash
+# Remove GPU cache
+rm -rf ~/.config/Cursor/GPUCache/
+
+# Remove code cache
+rm -rf ~/.config/Cursor/CachedData/
+
+# Try launching
+cursor --ozone-platform=wayland
+```
+
+### Option 3: Fresh Extensions (Keep Settings)
+
+```bash
+# Backup
+cp -r ~/.config/Cursor ~/.config/Cursor.backup
+
+# Remove extensions but keep settings
+rm -rf ~/.config/Cursor/extensions/
+
+# Try launching (extensions will reinstall)
+cursor --ozone-platform=wayland
+```
+
+### Option 4: Selective Settings Reset
+
+```bash
+# Backup
+cp -r ~/.config/Cursor ~/.config/Cursor.backup
+
+# Copy settings.json to safe place
+cp ~/.config/Cursor/User/settings.json ~/cursor-settings-backup.json
+
+# Remove problematic UI state
+rm -rf ~/.config/Cursor/User/globalStorage/
+rm -rf ~/.config/Cursor/User/workspaceStorage/
+
+# Restore settings
+cp ~/cursor-settings-backup.json ~/.config/Cursor/User/settings.json
+
+# Try launching
+cursor --ozone-platform=wayland
+```
+
+### Option 5: Full Profile Reset (Last Resort)
+
+Only if nothing else works:
+
+```bash
+# Full backup
+mv ~/.config/Cursor ~/.config/Cursor.broken-$(date +%Y%m%d)
+
+# Launch fresh
+cursor --ozone-platform=wayland
+
+# Then manually restore settings/keybindings from backup
+```
+
+## Advanced Debugging
+
+### Check Wayland Socket
+
+```bash
+# Verify Wayland display
+echo $WAYLAND_DISPLAY
+
+# Check socket permissions
+ls -la $XDG_RUNTIME_DIR/$WAYLAND_DISPLAY
+
+# Test with weston-info
+weston-info | head -20
+```
+
+### Check Niri Compositor
+
+Since you're using niri:
+
+```bash
+# Check niri is running
+ps aux | grep niri
+
+# Check niri IPC
+niri msg outputs
+
+# Check niri version
+niri --version
+```
+
+### Session Type Check
+
+Your logs showed `XDG_SESSION_TYPE=tty`. This might be the issue:
+
+```bash
+# Check session type
+echo $XDG_SESSION_TYPE
+# Should be "wayland", not "tty"
+
+# If wrong, ensure niri session is properly configured
+# Check /etc/greetd/environments
+cat /etc/greetd/environments
+```
+
+### Cursor Launch Script
+
+Check how Cursor is being launched:
+
+```bash
+# Find the Cursor wrapper
+which cursor
+cat $(which cursor)
+
+# Check for any custom launch flags in your shell config
+grep -r "cursor" ~/.zshrc ~/.bashrc ~/.config/fish/
+```
+
+## NixOS-Specific Checks
+
+### Verify Cursor Package
+
+```bash
+# Check Cursor derivation
+nix-store -q --references $(which cursor) | grep cursor
+
+# Verify Ozone flags are set system-wide
+echo $NIXOS_OZONE_WL
+# Should be "1"
+```
+
+### Check Home Manager Config
+
+```bash
+cd /etc/nixos/.dotfiles
+
+# Check for Cursor configuration
+grep -r "cursor" modules/apps/
+grep -r "Cursor" home/
+```
+
+From your sliceanddice config, Cursor is enabled via:
+```nix
+dendritic.apps.cursor.enable = true;
+```
+
+### Rebuild if Needed
+
+Only if you suspect the NixOS configuration:
+
+```bash
+cd /etc/nixos/.dotfiles
+sudo nixos-rebuild switch --flake .#sliceanddice
+```
+
+## Automated Fix Script
+
+```bash
+#!/usr/bin/env bash
+# Quick Cursor fix script
+
+echo "=== Cursor Multi-Instance Fix ==="
+echo ""
+
+echo "Step 1: Killing all Cursor processes..."
+pkill -9 cursor 2>/dev/null || true
+pkill -9 Cursor 2>/dev/null || true
+sleep 2
+
+echo "Step 2: Checking for remaining processes..."
+if pgrep -i cursor >/dev/null; then
+    echo "⚠️  Cursor processes still running:"
+    ps aux | grep -i [c]ursor
+else
+    echo "✓ All Cursor processes killed"
+fi
+echo ""
+
+echo "Step 3: Testing clean launch with Wayland..."
+cursor --ozone-platform=wayland &
+CURSOR_PID=$!
+sleep 3
+
+if ps -p $CURSOR_PID > /dev/null; then
+    echo "✓ Cursor launched successfully!"
+    echo "  PID: $CURSOR_PID"
+    echo ""
+    echo "If the window appeared, you're done!"
+    echo "If not, press Ctrl+C and run the manual tests."
+else
+    echo "❌ Cursor failed to launch"
+    echo "   Try the manual debugging steps in FIX-CURSOR-MULTIPLE-INSTANCES.md"
+fi
+```
+
+Save this as `fix-cursor-quick.sh` and run it.
+
+## Prevention
+
+To avoid this issue in the future:
+
+1. **Always use native Wayland**:
+   ```bash
+   # Add to your shell rc file
+   alias cursor='cursor --ozone-platform=wayland'
+   ```
+
+2. **Clean shutdown**: Don't force-kill Cursor unless necessary
+   
+3. **Monitor stuck processes**:
+   ```bash
+   # Before launching, check if already running
+   pgrep cursor && echo "Cursor already running!"
+   ```
+
+4. **Regular cache cleanup**:
+   ```bash
+   # Once a month
+   rm -rf ~/.config/Cursor/GPUCache/
+   rm -rf ~/.config/Cursor/CachedData/
+   ```
+
+## Summary of the Fix Path
+
+```
+1. Kill all Cursor processes
+   ↓ Works? → Done!
+   ↓ Fails? ↓
+   
+2. Launch without extensions  
+   ↓ Works? → Disable problematic extension
+   ↓ Fails? ↓
+   
+3. Launch without GPU
+   ↓ Works? → Fix Mesa/GPU configuration
+   ↓ Fails? ↓
+   
+4. Launch with fresh profile
+   ↓ Works? → Repair original profile
+   ↓ Fails? → Advanced debugging needed
+```
+
+## Most Likely Solution
+
+Based on the symptoms, **the most likely fix** is simply:
+
+```bash
+# Kill stuck process
+pkill -9 cursor
+
+# Relaunch with proper Wayland
+cursor --ozone-platform=wayland
+```
+
+The isolated launch proved Cursor works fine; the issue was the stuck existing process or corrupted window state.
+
+## What NOT to Do
+
+- ❌ Don't reinstall Cursor (not needed)
+- ❌ Don't rebuild NixOS (not needed)  
+- ❌ Don't delete `~/.config/Cursor` without backup
+- ❌ Don't delete `~/.cursor` without backup
+
+## Need More Help?
+
+If none of these work, collect debug info:
+
+```bash
+# Run Cursor with verbose logging
+cursor --ozone-platform=wayland --enable-logging --v=1 2>&1 | tee ~/cursor-debug.log
+
+# Then check the log
+grep -i "error\|failed\|denied" ~/cursor-debug.log
+```
diff --git a/FIX-TTY1-GREETER.md b/FIX-TTY1-GREETER.md
new file mode 100644
index 00000000..6f9bf19e
--- /dev/null
+++ b/FIX-TTY1-GREETER.md
@@ -0,0 +1,296 @@
+# Fix: TTY1 Blank Cursor / Greeter Not Starting
+
+## Problem
+TTY1 shows a blank blinking cursor instead of the graphical greeter (gtkgreet running in sway kiosk mode).
+
+## Root Cause Analysis
+
+Based on the niri configuration in `modules/apps/niri.nix`, the login flow is:
+1. **greetd** service starts on boot
+2. **sway** launches as a minimal kiosk compositor  
+3. **gtkgreet** runs inside sway to show the login UI
+4. After login, **niri** starts as the main desktop compositor
+
+**Blank cursor** means one of these steps is failing (most likely #2 or #3).
+
+## Quick Diagnosis
+
+### From TTY2 (Ctrl+Alt+F2)
+
+```bash
+# Login to TTY2
+# Then check what's failing:
+
+# 1. Check greetd service
+sudo systemctl status greetd
+
+# 2. Check recent logs
+sudo journalctl -u greetd -n 50
+
+# 3. Check for sway processes
+ps aux | grep sway
+
+# 4. Check graphics
+ls -la /dev/dri/
+```
+
+### Using the Diagnostic Script
+
+```bash
+cd /etc/nixos/.dotfiles
+sudo bash diagnose-tty1-greeter.sh
+```
+
+## Common Causes & Fixes
+
+### 1. Sway Kiosk Config Missing/Broken
+
+**Symptom**: greetd tries to start sway but sway fails immediately
+
+**Check**:
+```bash
+# Look for the sway greeter config in system closure
+ls -la /nix/store/*greetd-sway-gtkgreet*
+```
+
+**Fix**: The config is generated from `modules/apps/niri.nix`. Check if `swayGreeterConfig` is being created correctly.
+
+### 2. Graphics Driver Issues (NVIDIA + Intel Hybrid)
+
+**Symptom**: sway can't access GPU, DRM errors in journal
+
+**Check**:
+```bash
+# Verify drivers are loaded
+lsmod | grep -E "nvidia|i915"
+
+# Check DRM devices
+ls -la /dev/dri/
+
+# Look for DRM errors
+journalctl -b | grep -i drm
+```
+
+**Fix for NVIDIA issues**:
+- The sliceanddice config uses NVIDIA open drivers + Intel hybrid
+- sway needs access to the Intel iGPU (not NVIDIA for the greeter)
+- Ensure `hardware.nvidia.modesetting.enable = true`
+- Check that sway isn't trying to use NVIDIA
+
+### 3. gtkgreet Binary Missing or Wrong Path
+
+**Symptom**: sway starts but gtkgreet command fails
+
+**Check**:
+```bash
+# Verify gtkgreet exists
+which gtkgreet
+gtkgreet --version
+
+# Check the exact command greetd is trying to run
+sudo journalctl -u greetd | grep -i "exec\|command"
+```
+
+**Fix**: Rebuild to ensure gtkgreet is in the closure:
+```bash
+cd /etc/nixos/.dotfiles
+sudo nixos-rebuild switch --flake .#sliceanddice
+```
+
+### 4. CSS File Missing
+
+**Symptom**: gtkgreet starts but crashes due to missing stylesheet
+
+**Check**:
+```bash
+ls -la /etc/greetd/gtkgreet.css
+```
+
+**Fix**: The CSS is generated from `_gtk-auth-style.nix` via `environment.etc."greetd/gtkgreet.css"`. Rebuild to regenerate it.
+
+### 5. Wayland Socket Permissions
+
+**Symptom**: sway starts but can't create Wayland socket
+
+**Check**:
+```bash
+ls -la /run/user/*/wayland-*
+sudo -u greeter ls -la $XDG_RUNTIME_DIR 2>/dev/null || echo "No runtime dir"
+```
+
+**Fix**: Ensure the `greeter` user has proper runtime directory setup.
+
+### 6. Development vs Master Branch Mismatch
+
+**Important**: The development branch has a completely different auth setup!
+
+**Development branch** uses:
+- Runtime wallpaper selection
+- Dynamic CSS generation
+- `gtkgreet-auth` wrapper script
+- Avatar/profile photo integration
+- Dendritic appearance tool
+
+**Master branch** uses:
+- Static wallpaper from `stylix.image`
+- Static CSS from `_gtk-auth-style.nix`
+- Simple gtkgreet launch
+
+**If you're seeing issues**, check which branch you're on:
+```bash
+cd /etc/nixos/.dotfiles
+git branch --show-current
+```
+
+If on development, the auth system is more complex and may have additional failure points.
+
+## Step-by-Step Fix
+
+### Fix 1: Restart greetd
+
+```bash
+sudo systemctl restart greetd
+```
+
+Then switch back to TTY1 (Ctrl+Alt+F1) and see if greeter appears.
+
+### Fix 2: Rebuild from Stable Master
+
+```bash
+cd /etc/nixos/.dotfiles
+
+# Switch to master if needed
+git checkout master
+git pull origin master
+
+# Rebuild
+sudo nixos-rebuild switch --flake .#sliceanddice
+
+# Reboot
+sudo reboot
+```
+
+### Fix 3: Check and Fix Graphics
+
+```bash
+# If NVIDIA is causing issues, temporarily disable it for greeter
+# Edit hosts/nixos/sliceanddice/default.nix
+# Comment out the nvidia videoDrivers line:
+# services.xserver.videoDrivers = [ "nvidia" ];
+
+sudo nixos-rebuild switch --flake /etc/nixos/.dotfiles#sliceanddice
+sudo reboot
+```
+
+### Fix 4: Switch to Different Greeter (Emergency)
+
+If gtkgreet is broken, you can temporarily use autologin or a simpler greeter:
+
+```nix
+# In hosts/nixos/sliceanddice/default.nix
+# Temporarily disable niri greeter:
+dendritic.apps.niri.enable = false;
+
+# Add simple getty autologin:
+services.getty.autologinUser = "alex";
+```
+
+## Manual Testing
+
+To test the greeter command manually:
+
+```bash
+# Become the greeter user
+sudo -u greeter bash
+
+# Try running sway with the greeter config
+# (You'll need to find the config path first)
+sway --config /nix/store/*greetd-sway-gtkgreet*
+```
+
+## Checking Configuration
+
+### Expected Configuration on Master Branch
+
+From `modules/apps/niri.nix`:
+
+```nix
+swayGreeterConfig = pkgs.writeText "greetd-sway-gtkgreet" ''
+  # Minimal kiosk compositor for gtkgreet (desktop sway stays disabled).
+  exec "${gtkgreet}/bin/gtkgreet -l -s /etc/greetd/gtkgreet.css; ${pkgs.sway}/bin/swaymsg exit"
+  bindsym Mod4+shift+e exec ${pkgs.sway}/bin/swaynag \
+    -t warning \
+    -m 'Power?' \
+    -b 'Poweroff' 'systemctl poweroff' \
+    -b 'Reboot' 'systemctl reboot'
+'';
+
+services.greetd = {
+  enable = true;
+  settings.default_session = {
+    command = "${pkgs.sway}/bin/sway --config ${swayGreeterConfig}";
+    user = "greeter";
+  };
+};
+```
+
+This should be present if `dendritic.apps.niri.enable = true` in your sliceanddice config.
+
+## About "Shoji" Greeter
+
+You mentioned "Shoji-wm/Shoji greeter" - this is **not** in the current codebase. The actual setup uses:
+- **greetd** (login daemon)
+- **gtkgreet** (GTK-based greeter UI)
+- **sway** (Wayland compositor, kiosk mode only for login)
+- **niri** (main desktop compositor after login)
+
+If you're seeing references to "Shoji", this might be:
+1. A local modification not in git
+2. A misunderstanding of the stack
+3. An experimental branch
+
+## System Stack Verification
+
+```bash
+# On the actual sliceanddice machine:
+cd /etc/nixos/.dotfiles
+
+# Check configuration
+grep -r "shoji" . || echo "No Shoji found"
+grep -r "greetd" hosts/nixos/sliceanddice/
+grep -r "gtkgreet" modules/apps/niri.nix
+
+# Verify packages are installed
+nix-store -q --references /run/current-system | grep -E "(greetd|gtkgreet|sway)"
+```
+
+## Logs to Collect
+
+If the issue persists, collect these logs:
+
+```bash
+# greetd logs
+sudo journalctl -u greetd -n 100 > /tmp/greetd.log
+
+# Boot logs (graphics/wayland)
+sudo journalctl -b | grep -iE "(sway|wayland|drm|nvidia|i915)" > /tmp/graphics.log
+
+# System status
+systemctl status greetd > /tmp/greetd-status.log
+systemctl status getty@tty1 > /tmp/getty-status.log
+```
+
+## Next Steps
+
+1. Run `diagnose-tty1-greeter.sh` and review output
+2. Check which branch you're on (master vs development)
+3. Check greetd logs for specific error messages
+4. Verify graphics drivers are working
+5. If all else fails, boot previous generation or use autologin temporarily
+
+## Prevention
+
+- **Test display manager changes in VM first**
+- **Keep old generations**: Don't set `boot.loader.systemd-boot.configurationLimit` too low
+- **Stick to one branch**: Don't mix master and development features
+- **Check logs after every rebuild**: `sudo journalctl -u greetd -f`
diff --git a/diagnose-tty1-greeter.sh b/diagnose-tty1-greeter.sh
new file mode 100755
index 00000000..89a3e377
--- /dev/null
+++ b/diagnose-tty1-greeter.sh
@@ -0,0 +1,157 @@
+#!/usr/bin/env bash
+# Diagnostic for TTY1 greeter issues (blank blinking cursor)
+# Run this as root to diagnose greeter/display manager problems
+
+set -euo pipefail
+
+echo "=== TTY1 / Greeter Diagnostic ==="
+echo "Timestamp: $(date)"
+echo ""
+
+echo "=== Greeter Service Status ==="
+echo "--- greetd status ---"
+systemctl status greetd.service --no-pager -l || true
+echo ""
+
+echo "--- greetd logs (last 50 lines) ---"
+journalctl -u greetd.service -n 50 --no-pager || true
+echo ""
+
+echo "=== Display Manager Check ==="
+echo "Active display managers:"
+systemctl list-units --type=service --state=running | grep -iE "(gdm|sddm|lightdm|greetd|ly)" || echo "None found"
+echo ""
+
+echo "=== TTY1 Status ==="
+echo "--- getty@tty1 status ---"
+systemctl status getty@tty1.service --no-pager || true
+echo ""
+
+echo "=== Sway / Compositor Status ==="
+echo "Sway processes:"
+ps aux | grep -E "[s]way|[S]way" || echo "No sway processes"
+echo ""
+
+echo "Niri processes:"
+ps aux | grep "[n]iri" || echo "No niri processes"
+echo ""
+
+echo "=== Wayland / Display Environment ==="
+echo "WAYLAND_DISPLAY: ${WAYLAND_DISPLAY:-not set}"
+echo "DISPLAY: ${DISPLAY:-not set}"
+echo ""
+
+echo "Active Wayland compositors:"
+ps aux | grep -E "[w]eston|[s]way|[n]iri|[h]yprland|[w]ay.*fire" || echo "None found"
+echo ""
+
+echo "=== gtkgreet Configuration ==="
+if [ -f /etc/greetd/config.toml ]; then
+    echo "--- /etc/greetd/config.toml ---"
+    cat /etc/greetd/config.toml
+else
+    echo "⚠️  /etc/greetd/config.toml NOT FOUND"
+fi
+echo ""
+
+if [ -f /etc/greetd/gtkgreet.css ]; then
+    echo "✓ /etc/greetd/gtkgreet.css exists ($(wc -l < /etc/greetd/gtkgreet.css) lines)"
+else
+    echo "⚠️  /etc/greetd/gtkgreet.css NOT FOUND"
+fi
+echo ""
+
+if [ -d /etc/greetd ]; then
+    echo "--- /etc/greetd contents ---"
+    ls -la /etc/greetd/
+else
+    echo "⚠️  /etc/greetd directory NOT FOUND"
+fi
+echo ""
+
+echo "=== Sway Greeter Config ==="
+sway_config=$(nix-store -q --references /run/current-system 2>/dev/null | grep -i "greetd-sway" | head -1 || echo "")
+if [ -n "$sway_config" ]; then
+    echo "Found sway greeter config: $sway_config"
+    if [ -f "$sway_config" ]; then
+        echo "--- Contents ---"
+        cat "$sway_config" 2>/dev/null || echo "Cannot read"
+    fi
+else
+    echo "⚠️  Sway greeter config not found in system closure"
+fi
+echo ""
+
+echo "=== Recent Boot Errors (grep for sway/greeter/wayland) ==="
+journalctl -b --no-pager -p err | grep -iE "(sway|greet|wayland|compositor|drm|gpu)" || echo "No relevant errors"
+echo ""
+
+echo "=== Graphics/DRM Status ==="
+echo "--- DRM devices ---"
+ls -la /dev/dri/ 2>/dev/null || echo "/dev/dri not found"
+echo ""
+
+echo "--- Graphics drivers loaded ---"
+lsmod | grep -iE "(nvidia|nouveau|i915|amdgpu|radeon)" || echo "No graphics modules"
+echo ""
+
+echo "=== NixOS Configuration ==="
+if [ -d /etc/nixos/.dotfiles ]; then
+    cd /etc/nixos/.dotfiles
+    echo "Git branch: $(git branch --show-current 2>/dev/null || echo 'unknown')"
+    echo "Last commit: $(git log -1 --oneline 2>/dev/null || echo 'unknown')"
+    echo ""
+    
+    echo "Checking niri module config..."
+    if grep -q "dendritic.apps.niri.enable = true" hosts/nixos/sliceanddice/default.nix 2>/dev/null; then
+        echo "✓ niri is enabled"
+    else
+        echo "⚠️  niri might not be enabled"
+    fi
+    
+    if grep -q "services.greetd" modules/apps/niri.nix 2>/dev/null; then
+        echo "✓ greetd configuration found in niri.nix"
+    else
+        echo "⚠️  greetd configuration not found"
+    fi
+fi
+echo ""
+
+echo "=== Package Versions ==="
+echo "gtkgreet: $(gtkgreet --version 2>/dev/null || echo 'not in PATH')"
+echo "sway: $(sway --version 2>/dev/null || echo 'not in PATH')"
+echo "niri: $(niri --version 2>/dev/null || echo 'not in PATH')"
+echo ""
+
+echo "=== Common Issues & Solutions ==="
+echo ""
+echo "Issue: Blank cursor on TTY1"
+echo "Possible causes:"
+echo "  1. Sway failing to start (check sway logs above)"
+echo "  2. gtkgreet crashing immediately (check greetd logs)"
+echo "  3. Missing/broken sway config for greeter"
+echo "  4. Graphics driver issues (NVIDIA/Intel hybrid)"
+echo "  5. Wayland socket permissions"
+echo ""
+echo "Quick fixes:"
+echo "  1. Switch to TTY2: Ctrl+Alt+F2, login, check logs"
+echo "  2. Restart greetd: systemctl restart greetd"
+echo "  3. Check sway manually: sudo -u greeter sway --config /path/to/greeter/config"
+echo "  4. Rebuild config: nixos-rebuild switch --flake /etc/nixos/.dotfiles#sliceanddice"
+echo ""
+
+echo "=== Suggested Next Steps ==="
+echo ""
+if systemctl is-failed greetd.service >/dev/null 2>&1; then
+    echo "❌ greetd service FAILED - this is the problem!"
+    echo "   → Check 'journalctl -u greetd.service' output above"
+    echo "   → Common fix: rebuild NixOS config"
+elif ! systemctl is-active greetd.service >/dev/null 2>&1; then
+    echo "❌ greetd service NOT RUNNING"
+    echo "   → Try: systemctl start greetd"
+else
+    echo "✓ greetd service is running"
+    echo "  → Check if sway process is stuck"
+    echo "  → Check Wayland socket: ls -la /run/user/*/wayland-*"
+fi
+echo ""

From db9108500b4610c86c75824c7ead93a1047b99f6 Mon Sep 17 00:00:00 2001
From: Cursor Agent <cursoragent@cursor.com>
Date: Wed, 11 Nov 2026 17:59:09 +0000
Subject: [PATCH 8/8] docs: add comprehensive current issues summary

Combines emergency mode, TTY1 greeter, and Cursor multi-instance issues
into one reference document with quick fixes and documentation index.

Co-authored-by: Alex Spaulding <aspauldingcode@gmail.com>
---
 CURRENT-ISSUES-SUMMARY.md | 240 ++++++++++++++++++++++++++++++++++++++
 1 file changed, 240 insertions(+)
 create mode 100644 CURRENT-ISSUES-SUMMARY.md

diff --git a/CURRENT-ISSUES-SUMMARY.md b/CURRENT-ISSUES-SUMMARY.md
new file mode 100644
index 00000000..8bf429d4
--- /dev/null
+++ b/CURRENT-ISSUES-SUMMARY.md
@@ -0,0 +1,240 @@
+# Current Issues Summary - sliceanddice
+
+**Date**: November 11, 2026  
+**System**: sliceanddice (NixOS, niri desktop)  
+**Status**: Two active issues, both diagnosed with solutions ready
+
+---
+
+## Issue #1: TTY1 Blank Cursor / Greeter Not Starting
+
+### Status: 🔴 CRITICAL
+**What**: TTY1 shows blank blinking cursor instead of graphical login (gtkgreet)  
+**Impact**: Cannot log in graphically  
+**Root Cause**: greetd → sway → gtkgreet chain is broken
+
+### Quick Fix:
+```bash
+# From TTY2 (Ctrl+Alt+F2)
+sudo systemctl restart greetd
+```
+
+### Alternative Fix:
+```bash
+# Boot previous generation from systemd-boot menu
+# Or rebuild from stable master branch
+cd /etc/nixos/.dotfiles
+git checkout master
+sudo nixos-rebuild switch --flake .#sliceanddice
+```
+
+### Documentation:
+- **FIX-TTY1-GREETER.md** - Complete troubleshooting guide
+- **diagnose-tty1-greeter.sh** - Diagnostic script
+
+### Technical Details:
+The login flow is:
+1. greetd service starts
+2. sway launches as kiosk compositor  
+3. gtkgreet shows login UI inside sway
+4. After login, niri starts as desktop
+
+One of these steps is failing. Most likely sway or gtkgreet.
+
+### Important Note:
+You mentioned "Shoji-wm/Shoji greeter" but this is **not** in the codebase. The actual stack is:
+- greetd (service)
+- sway (kiosk compositor)
+- gtkgreet (login UI)
+- niri (desktop)
+
+If you're seeing "Shoji" references, check for local modifications.
+
+---
+
+## Issue #2: Cursor Multiple Instances / Display Issues
+
+### Status: 🟡 MODERATE  
+**What**: Cursor running but not displaying windows properly  
+**Impact**: Cannot use Cursor IDE  
+**Root Cause**: Stuck Cursor process or corrupted window state
+
+### Quick Fix:
+```bash
+# Kill all Cursor processes
+pkill -9 cursor
+
+# Relaunch with Wayland
+cursor --ozone-platform=wayland
+```
+
+### Key Findings:
+| Finding | Meaning | Status |
+|---------|---------|--------|
+| Isolated profile works | Fresh user-data launches fine | ✅ Confirmed |
+| Existing instance stuck | PID 23387 not showing windows | ✅ Confirmed |
+| X11 backend active | Using XWayland instead of native Wayland | ✅ Confirmed |
+| Mesa error | GBM driver load failure | ⚠️ Unknown impact |
+
+### Why This Isn't a Reinstall:
+- Cursor works (isolated launch proved it)
+- NixOS config is fine
+- **Problem is in `~/.config/Cursor` state**
+
+### Progressive Fixes:
+1. **Kill and relaunch** (most likely fix)
+2. **Disable extensions** if #1 fails
+3. **Disable GPU** if #2 fails  
+4. **Fresh profile** if #3 fails
+5. **Repair original profile** selectively
+
+### Documentation:
+- **FIX-CURSOR-MULTIPLE-INSTANCES.md** - Complete debugging guide
+
+### Don't Do This:
+- ❌ Don't reinstall Cursor
+- ❌ Don't rebuild NixOS
+- ❌ Don't delete config without backup
+
+---
+
+## Combined Recovery Plan
+
+### Priority Order:
+
+**1. Fix TTY1 First** (you need graphical login)
+   - Use TTY2 to diagnose
+   - Restart greetd service
+   - Or boot previous generation
+
+**2. Then Fix Cursor** (after you can log in)
+   - Kill stuck processes
+   - Relaunch with Wayland flags
+   - Test progressively if needed
+
+### Emergency Access:
+
+If you can't log in graphically:
+
+```bash
+# From TTY2
+# Temporarily enable autologin
+sudo nano /etc/nixos/.dotfiles/hosts/nixos/sliceanddice/default.nix
+
+# Add:
+services.getty.autologinUser = "alex";
+
+# Rebuild
+sudo nixos-rebuild switch --flake /etc/nixos/.dotfiles#sliceanddice
+
+# Reboot
+sudo reboot
+```
+
+---
+
+## System Context
+
+### Hardware:
+- **Machine**: MSI laptop
+- **CPU**: Intel Tiger Lake
+- **GPU**: Intel iGPU + NVIDIA RTX 3050 Ti (hybrid)
+- **Display**: Internal panel on Intel iGPU
+- **Storage**: Samsung 870 EVO 500GB
+
+### Software Stack:
+- **OS**: NixOS (system.stateVersion 24.11)
+- **Compositor**: niri (scrollable-tiling Wayland)
+- **Greeter**: gtkgreet via sway kiosk
+- **Service Manager**: greetd
+- **Kernel**: linuxPackages_latest
+
+### Current Branch Status:
+- **Master**: Stable, working configuration
+- **Development**: New features, requires additional files
+- **Emergency mode issue**: Caused by branch mismatch (now documented)
+
+---
+
+## All Available Documentation
+
+| File | Purpose |
+|------|---------|
+| **CURRENT-ISSUES-SUMMARY.md** | This file - overview of both issues |
+| **FIX-TTY1-GREETER.md** | TTY1/greeter troubleshooting |
+| **diagnose-tty1-greeter.sh** | TTY1 diagnostic script |
+| **FIX-CURSOR-MULTIPLE-INSTANCES.md** | Cursor issues troubleshooting |
+| **START_HERE.md** | Emergency mode recovery quick start |
+| **README-EMERGENCY.md** | Emergency recovery hub |
+| **QUICKSTART.md** | Fast emergency recovery |
+| **SOLUTION.md** | Full emergency mode explanation |
+| **EMERGENCY_RECOVERY.md** | Technical recovery procedures |
+| **diagnose-emergency.sh** | Emergency mode diagnostic |
+| **fix-emergency-auto.sh** | Automated emergency fix |
+
+---
+
+## Quick Command Reference
+
+### Diagnostics:
+```bash
+# TTY1/Greeter
+sudo bash diagnose-tty1-greeter.sh
+
+# Emergency Mode
+sudo bash diagnose-emergency.sh
+
+# Cursor
+cursor --ozone-platform=wayland --enable-logging --v=1
+```
+
+### Fixes:
+```bash
+# TTY1
+sudo systemctl restart greetd
+
+# Cursor
+pkill -9 cursor && cursor --ozone-platform=wayland
+
+# Emergency Mode
+sudo bash fix-emergency-auto.sh
+```
+
+### Logs:
+```bash
+# Greeter
+sudo journalctl -u greetd -n 50
+
+# System
+sudo journalctl -b | grep -iE "(error|fail)"
+
+# Graphics
+sudo journalctl -b | grep -iE "(drm|nvidia|i915)"
+```
+
+---
+
+## Next Steps
+
+1. **Immediate**: Fix TTY1 to get graphical login working
+2. **Then**: Fix Cursor once logged in
+3. **Later**: Review and merge emergency mode documentation (PR #188)
+4. **Future**: Decide on master vs development branch strategy
+
+---
+
+## Notes
+
+- Both issues are **not** NixOS installation problems
+- Both have **clear diagnostic paths**
+- Both have **progressive fixes** (start simple, get more complex)
+- All diagnostics and fixes are now documented and committed
+- PR #188 contains all emergency recovery documentation
+
+---
+
+**Bottom Line**:
+1. TTY1 issue: Service/compositor problem → restart greetd or boot old generation
+2. Cursor issue: Stuck process → kill and relaunch with Wayland
+
+Both are fixable without reinstalling anything!