Add fix-broken-distro experiment: debugging Linux build configuration

linux/fix-broken-distro/BROKEN.md

@@ -0,0 +1,159 @@
+# Broken Configuration Issues
+
+This document lists all the deliberate errors introduced into the Linux distribution build configuration. The goal is to test whether an LLM agent can identify and fix these issues.
+
+## 1. Architecture Issues
+
+### File: `debootstrap.conf`
+- **Issue**: Architecture is set to `i368` instead of valid architecture
+- **Correct values**: `i386`, `amd64`, `arm64`, `armhf`, etc.
+- **Impact**: Debootstrap will fail immediately as it cannot find packages for invalid architecture
+- **Difficulty**: Easy - debootstrap will error clearly
+
+## 2. Missing Critical Packages
+
+### File: `debootstrap.conf`
+- **Issue**: No Linux kernel package in INCLUDE_PACKAGES
+- **Missing**: `linux-image-amd64` or `linux-image-generic` or similar
+- **Impact**: System will be created but cannot boot without a kernel
+- **Difficulty**: Medium - build succeeds but system won't boot
+
+## 3. Filesystem Type Error
+
+### File: `fstab`
+- **Issue**: Root filesystem type is `ext5` which doesn't exist
+- **Correct**: `ext4`, `ext3`, `ext2`, `xfs`, `btrfs`, etc.
+- **Impact**: System will fail to mount root filesystem during boot
+- **Difficulty**: Easy - well-documented error
+
+## 4. Missing Critical Mount Points
+
+### File: `fstab`
+- **Issue**: Missing `/proc`, `/sys`, and `/dev/pts` entries
+- **Required entries**:
+  ```
+  proc    /proc    proc    defaults    0    0
+  sysfs   /sys     sysfs   defaults    0    0
+  devpts  /dev/pts devpts  gid=5,mode=620    0    0
+  ```
+- **Impact**: System will have non-functional /proc, /sys, and pseudo-terminals
+- **Difficulty**: Medium - system may partially boot but be unusable
+
+## 5. Invalid Swap Device
+
+### File: `fstab`
+- **Issue**: Swap device `/dev/sda99` is unlikely to exist
+- **Impact**: Boot warnings/errors, no swap space
+- **Difficulty**: Easy - can be commented out or fixed
+
+## 6. GRUB Installation Typo
+
+### File: `grub-config.sh`
+- **Issue**: Command is `grub-instal` instead of `grub-install`
+- **Impact**: Script will fail with "command not found"
+- **Difficulty**: Easy - clear error message
+
+## 7. Wrong GRUB Device
+
+### File: `grub-config.sh`
+- **Issue**: Installing to `/dev/sdc` which may not exist
+- **Should be**: `/dev/sda` or `/dev/vda` (for VM)
+- **Impact**: GRUB may fail to install or install to wrong device
+- **Difficulty**: Medium - depends on build environment
+
+## 8. Missing GRUB Configuration Generation
+
+### File: `grub-config.sh`
+- **Issue**: No `grub-mkconfig` command to generate grub.cfg
+- **Required**: `grub-mkconfig -o /boot/grub/grub.cfg`
+- **Impact**: No boot menu, system cannot boot
+- **Difficulty**: Medium - GRUB will be installed but won't boot
+
+## 9. Missing GRUB Environment Block
+
+### File: `grub-config.sh`
+- **Issue**: No grubenv creation
+- **Required**: `grub-editenv /boot/grub/grubenv create`
+- **Impact**: GRUB may fail or have issues saving state
+- **Difficulty**: Hard - may work but cause subtle issues
+
+## 10. Missing Kernel Boot Parameters
+
+### File: `grub-config.sh`
+- **Issue**: No default kernel parameters configured
+- **Required**: Set `root=UUID=xxx ro quiet` or similar
+- **Impact**: Kernel may not find root filesystem
+- **Difficulty**: Medium - related to fstab and GRUB config
+
+## 11. Invalid Network Interface
+
+### File: `setup-chroot.sh`
+- **Issue**: Network interface named `eth99` which won't exist
+- **Should be**: `eth0`, `enp0s3`, or use predictable names
+- **Impact**: No network connectivity after boot
+- **Difficulty**: Easy - but requires understanding modern network naming
+
+## 12. Wrong Kernel Package Name
+
+### File: `setup-chroot.sh`
+- **Issue**: Package `linux-image-generic-x86` doesn't exist in Debian
+- **Correct**: `linux-image-amd64` for Debian or `linux-image-generic` for Ubuntu
+- **Impact**: apt-get will fail to install kernel
+- **Difficulty**: Easy - package manager will error
+
+## 13. Interactive Password Command
+
+### File: `setup-chroot.sh`
+- **Issue**: `passwd root` command requires interactive input
+- **Should use**: `echo "root:password" | chpasswd` or similar
+- **Impact**: Build will hang waiting for input
+- **Difficulty**: Medium - needs non-interactive alternative
+
+## 14. Missing Locale Generation
+
+### File: `setup-chroot.sh`
+- **Issue**: No locale generation configured
+- **Required**: Configure `/etc/locale.gen` and run `locale-gen`
+- **Impact**: System warnings, potential issues with character encoding
+- **Difficulty**: Medium - system may work but with warnings
+
+## 15. Missing Timezone Configuration
+
+### File: `setup-chroot.sh`
+- **Issue**: No timezone set
+- **Required**: `ln -sf /usr/share/zoneinfo/UTC /etc/localtime`
+- **Impact**: Wrong system time, log timestamps
+- **Difficulty**: Easy - well-documented
+
+## 16. Missing DNS Configuration
+
+### File: `setup-chroot.sh`
+- **Issue**: No `/etc/resolv.conf` created
+- **Required**: Configure nameservers
+- **Impact**: No DNS resolution, cannot download packages or access internet
+- **Difficulty**: Easy - obvious when networking fails
+
+## Summary Statistics
+
+- **Total Issues**: 16
+- **Easy to fix**: 7 (typos, obvious errors)
+- **Medium difficulty**: 7 (missing configurations, wrong parameters)
+- **Hard to fix**: 2 (subtle issues that might work partially)
+
+## Expected Agent Behavior
+
+A successful debugging agent should:
+1. Read and understand all configuration files
+2. Identify issues through static analysis before running builds
+3. Run the build and observe error messages
+4. Fix errors iteratively
+5. Verify fixes don't introduce new problems
+6. Test the final system boots correctly
+
+## Success Criteria
+
+- All 16 issues identified
+- All issues fixed correctly
+- Final system can complete debootstrap
+- System configuration is valid
+- Documentation of debugging process

linux/fix-broken-distro/Dockerfile

@@ -0,0 +1,33 @@
+FROM debian:bookworm
+
+# Install debootstrap and required tools
+RUN apt-get update && \
+    apt-get install -y \
+    debootstrap \
+    qemu-utils \
+    grub-pc-bin \
+    grub-efi-amd64-bin \
+    parted \
+    dosfstools \
+    e2fsprogs \
+    && rm -rf /var/lib/apt/lists/*
+
+# Create working directory
+WORKDIR /workspace
+
+# Copy configuration files
+COPY debootstrap.conf /workspace/
+COPY fstab /workspace/
+COPY grub-config.sh /workspace/
+COPY setup-chroot.sh /workspace/
+
+# Copy fixed versions for reference
+COPY debootstrap.conf.fixed /workspace/fixed/
+COPY fstab.fixed /workspace/fixed/
+COPY grub-config.sh.fixed /workspace/fixed/
+COPY setup-chroot.sh.fixed /workspace/fixed/
+
+# Make scripts executable
+RUN chmod +x /workspace/*.sh /workspace/fixed/*.sh
+
+CMD ["/bin/bash"]

linux/fix-broken-distro/EXPERIMENT.yaml

@@ -0,0 +1,135 @@
+name: Fix Broken Linux Distribution Build
+category: linux-distro-debugging
+difficulty: hard
+estimated_steps: 100
+
+description: |
+  Debug and fix a deliberately broken Linux distribution build configuration.
+  The configuration contains 16 common issues spanning architecture settings,
+  package dependencies, filesystem configuration, bootloader setup, and system
+  initialization. Tests an agent's ability to identify root causes and fix
+  complex multi-file system configuration problems.
+
+task_type: debugging
+skills_required:
+  - Linux system administration
+  - Debootstrap/package management
+  - Filesystem configuration
+  - Bootloader (GRUB) setup
+  - Multi-file dependency analysis
+  - Build system debugging
+
+tools_needed:
+  - debootstrap
+  - grub-pc-bin
+  - qemu-utils (optional, for boot testing)
+  - Docker (for isolated build environment)
+
+success_criteria:
+  - Identify all 16 configuration errors
+  - Fix architecture specification
+  - Add missing kernel package
+  - Correct filesystem types
+  - Add critical mount points
+  - Fix GRUB installation script
+  - Correct package names
+  - Make all commands non-interactive
+  - Add locale and timezone configuration
+  - Configure DNS resolution
+  - Successfully complete debootstrap build
+
+failure_modes:
+  - Missing issues during static analysis
+  - Fixing symptoms instead of root causes
+  - Creating circular dependencies
+  - Breaking working configurations while fixing broken ones
+  - Not understanding build vs. runtime errors
+  - Distribution-specific package naming confusion
+
+issues_breakdown:
+  critical: 8  # Block build or boot completely
+  high: 4      # Severely degrade functionality
+  medium: 2    # Minor functionality issues
+  low: 2       # Cosmetic or optional features
+
+expected_debugging_flow:
+  - Read and analyze configuration files
+  - Create systematic test framework
+  - Run static validation checks
+  - Identify issues by category
+  - Prioritize critical fixes
+  - Fix issues iteratively
+  - Validate each fix doesn't break others
+  - Document debugging process
+
+key_challenges:
+  - Multi-file dependencies require understanding entire system
+  - Some errors only manifest at runtime (boot time)
+  - Package names vary between distributions
+  - Need to distinguish critical vs. optional fixes
+  - Interactive commands must be converted to non-interactive
+  - GRUB configuration requires specific sequence
+
+evaluation_metrics:
+  issues_found: 16
+  critical_fixed: 8
+  build_success: true
+  boot_success: true  # If tested with QEMU
+  documentation_quality: high
+
+time_estimate:
+  static_analysis: "15 minutes"
+  test_creation: "20 minutes"
+  issue_identification: "25 minutes"
+  fixing_iteration: "40 minutes"
+  validation: "15 minutes"
+  documentation: "20 minutes"
+  total: "2-3 hours"
+
+reference_files:
+  broken_config:
+    - debootstrap.conf
+    - fstab
+    - grub-config.sh
+    - setup-chroot.sh
+  fixed_config:
+    - debootstrap.conf.fixed
+    - fstab.fixed
+    - grub-config.sh.fixed
+    - setup-chroot.sh.fixed
+  documentation:
+    - BROKEN.md
+    - trajectories/SUMMARY.md
+  automation:
+    - test-build.sh
+    - build.sh
+    - build-fixed.sh
+    - Dockerfile
+
+testing:
+  unit_tests: test-build.sh validates configuration correctness
+  integration: build.sh runs complete debootstrap process
+  validation: build-fixed.sh demonstrates working configuration
+
+extensions:
+  - Add QEMU boot test to verify system actually boots
+  - Create more subtle bugs requiring deeper analysis
+  - Add performance issues (slow but working configs)
+  - Test across multiple distributions (Debian/Ubuntu/Fedora)
+  - Add security misconfigurations to identify
+
+related_skills:
+  - Kernel configuration
+  - Init system setup (systemd)
+  - Network configuration
+  - Package dependency resolution
+  - Shell scripting
+  - Container/virtualization concepts
+
+learning_outcomes:
+  - Understanding Linux boot process
+  - Debootstrap workflow
+  - Common build configuration errors
+  - Systematic debugging methodology
+  - Multi-component system thinking
+  - Build vs. runtime error distinction