lldb
LLDB debugger skill for C/C++/Swift/Objective-C programs. Use when debugging with LLDB on macOS, FreeBSD, or Linux-clang environments, mapping GDB mental models to LLDB commands, using LLDB in Xcode or VS Code, or debugging Swift/Objective-C. Activates on queries about LLDB commands, GDB to LLDB migration, Apple platform debugging, LLDB Python scripting, or IDE-integrated debugging with clang-built binaries.
Works with
---
name: lldb
description: LLDB debugger skill for C/C++/Swift/Objective-C programs. Use when debugging with LLDB on macOS, FreeBSD, or Linux-clang environments, mapping GDB mental models to LLDB commands, using LLDB in Xcode or VS Code, or debugging Swift/Objective-C. Activates on queries about LLDB commands, GDB to LLDB migration, Apple platform debugging, LLDB Python scripting, or IDE-integrated debugging with clang-built binaries.
license: MIT
---
# LLDB
## Purpose
Guide agents through LLDB sessions and map existing GDB knowledge to LLDB. Covers command differences, Apple specifics, Python scripting, and IDE integration.
## Triggers
- "I'm on macOS and need to debug a C++ program"
- "How does LLDB differ from GDB?"
- "How do I do [GDB command] in LLDB?"
- "LLDB shows `<unavailable>` for variables"
- "How do I use LLDB in VS Code?"
- "How do I write an LLDB Python script?"
## Workflow
### 1. Start LLDB
```bash
lldb ./prog # load binary
lldb ./prog -- arg1 arg2 # with arguments
lldb -p 12345 # attach to PID
lldb -c core.1234 # load core dump
lldb ./prog core.1234 # binary + core
```
### 2. GDB → LLDB command map
Source: <https://lldb.llvm.org/use/map.html>
| GDB | LLDB | Notes |
|-----|------|-------|
| `run [args]` | `process launch [args]` / `r` | |
| `continue` | `process continue` / `c` | |
| `next` | `thread step-over` / `n` | |
| `step` | `thread step-in` / `s` | |
| `nexti` | `thread step-inst-over` / `ni` | |
| `stepi` | `thread step-inst` / `si` | |
| `finish` | `thread step-out` / `finish` | |
| `break main` | `breakpoint set -n main` / `b main` | |
| `break file.c:42` | `breakpoint set -f file.c -l 42` / `b file.c:42` | |
| `break *0x400abc` | `breakpoint set -a 0x400abc` / `b -a 0x400abc` | |
| `watch x` | `watchpoint set variable x` / `wa s v x` | |
| `print x` | `frame variable x` / `p x` | |
| `print/x x` | `p/x x` | |
| `info locals` | `frame variable` / `fr v` | |
| `info args` | `frame variable --arguments` | |
| `backtrace` | `thread backtrace` / `bt` | |
| `frame N` | `frame select N` / `f N` | |
| `info threads` | `thread list` | |
| `thread N` | `thread select N` | |
| `thread apply all bt` | `thread backtrace all` | |
| `x/10wx addr` | `memory read -s4 -fx -c10 addr` / `x/10xw addr` | |
| `set var = 42` | `expression var = 42` / `expr var = 42` | |
| `quit` | `quit` / `q` | |
### 3. Breakpoints
```lldb
# By name
b main
breakpoint set --name foo
breakpoint set --name foo --condition 'x > 0'
# By file:line
b file.c:42
breakpoint set --file file.c --line 42
# By address
b -a 0x100003f20
# By regex
breakpoint set --func-regex '^MyClass::'
# List
breakpoint list / br l
# Delete
breakpoint delete 2
# Disable/enable
breakpoint disable 1
breakpoint enable 1
# Commands on hit
breakpoint command add 1
> p x
> continue
> DONE
```
### 4. Inspect state
```lldb
# Print variable
p x
frame variable x
p *ptr
p arr[0]
# Print expression
expression x * 2 + 1
expr (int)sqrt(9.0)
# All locals
frame variable
fr v -a # include arguments
# Registers
register read
register read rip rsp
# Memory
memory read --size 4 --format x --count 10 0x7fff0000
x/10xw 0x7fff0000 # GDB-compatible syntax
# Type info
image lookup --type MyClass
type lookup MyClass
```
### 5. Watchpoints
```lldb
watchpoint set variable x # write watchpoint
watchpoint set variable -w read x # read watchpoint
watchpoint set variable -w read_write x
watchpoint set expression -- &x # by address
watchpoint list
watchpoint delete 1
```
### 6. Threads
```lldb
thread list
thread select 3
thread backtrace all
thread backtrace --count 5 # limit depth
# Per-thread stepping
thread step-over # step this thread only
```
### 7. macOS / Apple specifics
```lldb
# Symbol lookup in shared cache
image lookup --address 0x18ab12345
image lookup --name objc_msgSend
# Objective-C method breakpoint
b "-[NSArray objectAtIndex:]"
b "+[NSString stringWithFormat:]"
# Inspect Objective-C object
po myObject # print-object (calls -description)
po [arr count]
# Show loaded libraries
image list
image list -b # brief (names only)
```
### 8. VS Code integration
Install the `CodeLLDB` extension. `.vscode/launch.json`:
```json
{
"version": "0.2.0",
"configurations": [
{
"name": "Debug (lldb)",
"type": "lldb",
"request": "launch",
"program": "${workspaceFolder}/build/prog",
"args": [],
"cwd": "${workspaceFolder}",
"preLaunchTask": "build"
}
]
}
```
### 9. LLDB Python scripting
```python
import lldb
def print_all_threads(debugger, command, result, internal_dict):
target = debugger.GetSelectedTarget()
process = target.GetProcess()
for thread in process:
print(f"Thread {thread.GetIndexID()}: {thread.GetName()}")
for frame in thread:
print(f" {frame}")
def __lldb_init_module(debugger, internal_dict):
debugger.HandleCommand('command script add -f myscript.print_all_threads pthreads')
```
Load: `command script import /path/to/myscript.py`
For a full GDB↔LLDB command map, see [references/gdb-lldb-map.md](references/gdb-lldb-map.md).
## Related skills
- Use `skills/debuggers/gdb` for GDB workflows
- Use `skills/debuggers/core-dumps` for core dump analysis
- Use `skills/compilers/clang` for building with debug infoMore Debugging skills
diagnosing-bugs
mattpocock/skills
Diagnosis loop for hard bugs and performance regressions. Use when the user says "diagnose"/"debug this", or reports something broken/throwing/failing/slow.
explore-code
lllllllama/rigorpilot-skills
Rigor Improve implementation leaf skill for auditable candidate implementation in deep learning research repositories. Use when the researcher explicitly authorizes exploratory work on an isolated branch or worktree to transplant modules, adapt a backbone, add LoRA or adapter layers, replace a head, or stitch together meaningful low-risk migration ideas with rollback-aware records in `explore_outputs/`. Do not use for end-to-end exploration orchestration on top of `current_research`, trusted baseline reproduction, conservative debugging, environment setup, verified contribution claims, or default repository analysis.
safe-debug
lllllllama/rigorpilot-skills
Rigor Debug / Rigor Audit skill for deep learning research work. Use when the user pastes a traceback, terminal error, CUDA OOM, checkpoint load failure, shape mismatch, NaN loss symptom, or training failure and wants conservative diagnosis before any patching, with debug fixes clearly separated from research contributions. Do not use for broad refactoring, speculative adaptation, automatic exploratory patching, or general repository familiarization.

