Phase 1: Repository & Infrastructure Setup

- Renamed directories: codex-rs -> llmx-rs, codex-cli -> llmx-cli
- Updated package.json files:
  - Root: llmx-monorepo
  - CLI: @llmx/llmx
  - SDK: @llmx/llmx-sdk
- Updated pnpm workspace configuration
- Renamed binary: codex.js -> llmx.js
- Updated environment variables: CODEX_* -> LLMX_*
- Changed repository URLs to valknar/llmx

🤖 Generated with Claude Code
This commit is contained in:
Sebastian Krüger
2025-11-11 14:01:52 +01:00
parent 052b052832
commit f237fe560d
1151 changed files with 41 additions and 35 deletions

View File

@@ -0,0 +1,32 @@
[package]
edition = "2024"
name = "codex-linux-sandbox"
version = { workspace = true }
[[bin]]
name = "codex-linux-sandbox"
path = "src/main.rs"
[lib]
name = "codex_linux_sandbox"
path = "src/lib.rs"
[lints]
workspace = true
[target.'cfg(target_os = "linux")'.dependencies]
clap = { workspace = true, features = ["derive"] }
codex-core = { workspace = true }
landlock = { workspace = true }
libc = { workspace = true }
seccompiler = { workspace = true }
[target.'cfg(target_os = "linux")'.dev-dependencies]
tempfile = { workspace = true }
tokio = { workspace = true, features = [
"io-std",
"macros",
"process",
"rt-multi-thread",
"signal",
] }

View File

@@ -0,0 +1,8 @@
# codex-linux-sandbox
This crate is responsible for producing:
- a `codex-linux-sandbox` standalone executable for Linux that is bundled with the Node.js version of the Codex CLI
- a lib crate that exposes the business logic of the executable as `run_main()` so that
- the `codex-exec` CLI can check if its arg0 is `codex-linux-sandbox` and, if so, execute as if it were `codex-linux-sandbox`
- this should also be true of the `codex` multitool CLI

View File

@@ -0,0 +1,145 @@
use std::collections::BTreeMap;
use std::path::Path;
use std::path::PathBuf;
use codex_core::error::CodexErr;
use codex_core::error::Result;
use codex_core::error::SandboxErr;
use codex_core::protocol::SandboxPolicy;
use landlock::ABI;
use landlock::Access;
use landlock::AccessFs;
use landlock::CompatLevel;
use landlock::Compatible;
use landlock::Ruleset;
use landlock::RulesetAttr;
use landlock::RulesetCreatedAttr;
use seccompiler::BpfProgram;
use seccompiler::SeccompAction;
use seccompiler::SeccompCmpArgLen;
use seccompiler::SeccompCmpOp;
use seccompiler::SeccompCondition;
use seccompiler::SeccompFilter;
use seccompiler::SeccompRule;
use seccompiler::TargetArch;
use seccompiler::apply_filter;
/// Apply sandbox policies inside this thread so only the child inherits
/// them, not the entire CLI process.
pub(crate) fn apply_sandbox_policy_to_current_thread(
sandbox_policy: &SandboxPolicy,
cwd: &Path,
) -> Result<()> {
if !sandbox_policy.has_full_network_access() {
install_network_seccomp_filter_on_current_thread()?;
}
if !sandbox_policy.has_full_disk_write_access() {
let writable_roots = sandbox_policy
.get_writable_roots_with_cwd(cwd)
.into_iter()
.map(|writable_root| writable_root.root)
.collect();
install_filesystem_landlock_rules_on_current_thread(writable_roots)?;
}
// TODO(ragona): Add appropriate restrictions if
// `sandbox_policy.has_full_disk_read_access()` is `false`.
Ok(())
}
/// Installs Landlock file-system rules on the current thread allowing read
/// access to the entire file-system while restricting write access to
/// `/dev/null` and the provided list of `writable_roots`.
///
/// # Errors
/// Returns [`CodexErr::Sandbox`] variants when the ruleset fails to apply.
fn install_filesystem_landlock_rules_on_current_thread(writable_roots: Vec<PathBuf>) -> Result<()> {
let abi = ABI::V5;
let access_rw = AccessFs::from_all(abi);
let access_ro = AccessFs::from_read(abi);
let mut ruleset = Ruleset::default()
.set_compatibility(CompatLevel::BestEffort)
.handle_access(access_rw)?
.create()?
.add_rules(landlock::path_beneath_rules(&["/"], access_ro))?
.add_rules(landlock::path_beneath_rules(&["/dev/null"], access_rw))?
.set_no_new_privs(true);
if !writable_roots.is_empty() {
ruleset = ruleset.add_rules(landlock::path_beneath_rules(&writable_roots, access_rw))?;
}
let status = ruleset.restrict_self()?;
if status.ruleset == landlock::RulesetStatus::NotEnforced {
return Err(CodexErr::Sandbox(SandboxErr::LandlockRestrict));
}
Ok(())
}
/// Installs a seccomp filter that blocks outbound network access except for
/// AF_UNIX domain sockets.
fn install_network_seccomp_filter_on_current_thread() -> std::result::Result<(), SandboxErr> {
// Build rule map.
let mut rules: BTreeMap<i64, Vec<SeccompRule>> = BTreeMap::new();
// Helper insert unconditional deny rule for syscall number.
let mut deny_syscall = |nr: i64| {
rules.insert(nr, vec![]); // empty rule vec = unconditional match
};
deny_syscall(libc::SYS_connect);
deny_syscall(libc::SYS_accept);
deny_syscall(libc::SYS_accept4);
deny_syscall(libc::SYS_bind);
deny_syscall(libc::SYS_listen);
deny_syscall(libc::SYS_getpeername);
deny_syscall(libc::SYS_getsockname);
deny_syscall(libc::SYS_shutdown);
deny_syscall(libc::SYS_sendto);
deny_syscall(libc::SYS_sendmsg);
deny_syscall(libc::SYS_sendmmsg);
// NOTE: allowing recvfrom allows some tools like: `cargo clippy` to run
// with their socketpair + child processes for sub-proc management
// deny_syscall(libc::SYS_recvfrom);
deny_syscall(libc::SYS_recvmsg);
deny_syscall(libc::SYS_recvmmsg);
deny_syscall(libc::SYS_getsockopt);
deny_syscall(libc::SYS_setsockopt);
deny_syscall(libc::SYS_ptrace);
// For `socket` we allow AF_UNIX (arg0 == AF_UNIX) and deny everything else.
let unix_only_rule = SeccompRule::new(vec![SeccompCondition::new(
0, // first argument (domain)
SeccompCmpArgLen::Dword,
SeccompCmpOp::Ne,
libc::AF_UNIX as u64,
)?])?;
rules.insert(libc::SYS_socket, vec![unix_only_rule.clone()]);
rules.insert(libc::SYS_socketpair, vec![unix_only_rule]); // always deny (Unix can use socketpair but fine, keep open?)
let filter = SeccompFilter::new(
rules,
SeccompAction::Allow, // default allow
SeccompAction::Errno(libc::EPERM as u32), // when rule matches return EPERM
if cfg!(target_arch = "x86_64") {
TargetArch::x86_64
} else if cfg!(target_arch = "aarch64") {
TargetArch::aarch64
} else {
unimplemented!("unsupported architecture for seccomp filter");
},
)?;
let prog: BpfProgram = filter.try_into()?;
apply_filter(&prog)?;
Ok(())
}

View File

@@ -0,0 +1,14 @@
#[cfg(target_os = "linux")]
mod landlock;
#[cfg(target_os = "linux")]
mod linux_run_main;
#[cfg(target_os = "linux")]
pub fn run_main() -> ! {
linux_run_main::run_main();
}
#[cfg(not(target_os = "linux"))]
pub fn run_main() -> ! {
panic!("codex-linux-sandbox is only supported on Linux");
}

View File

@@ -0,0 +1,56 @@
use clap::Parser;
use std::ffi::CString;
use std::path::PathBuf;
use crate::landlock::apply_sandbox_policy_to_current_thread;
#[derive(Debug, Parser)]
pub struct LandlockCommand {
/// It is possible that the cwd used in the context of the sandbox policy
/// is different from the cwd of the process to spawn.
#[arg(long = "sandbox-policy-cwd")]
pub sandbox_policy_cwd: PathBuf,
#[arg(long = "sandbox-policy")]
pub sandbox_policy: codex_core::protocol::SandboxPolicy,
/// Full command args to run under landlock.
#[arg(trailing_var_arg = true)]
pub command: Vec<String>,
}
pub fn run_main() -> ! {
let LandlockCommand {
sandbox_policy_cwd,
sandbox_policy,
command,
} = LandlockCommand::parse();
if let Err(e) = apply_sandbox_policy_to_current_thread(&sandbox_policy, &sandbox_policy_cwd) {
panic!("error running landlock: {e:?}");
}
if command.is_empty() {
panic!("No command specified to execute.");
}
#[expect(clippy::expect_used)]
let c_command =
CString::new(command[0].as_str()).expect("Failed to convert command to CString");
#[expect(clippy::expect_used)]
let c_args: Vec<CString> = command
.iter()
.map(|arg| CString::new(arg.as_str()).expect("Failed to convert arg to CString"))
.collect();
let mut c_args_ptrs: Vec<*const libc::c_char> = c_args.iter().map(|arg| arg.as_ptr()).collect();
c_args_ptrs.push(std::ptr::null());
unsafe {
libc::execvp(c_command.as_ptr(), c_args_ptrs.as_ptr());
}
// If execvp returns, there was an error.
let err = std::io::Error::last_os_error();
panic!("Failed to execvp {}: {err}", command[0].as_str());
}

View File

@@ -0,0 +1,6 @@
/// Note that the cwd, env, and command args are preserved in the ultimate call
/// to `execv`, so the caller is responsible for ensuring those values are
/// correct.
fn main() -> ! {
codex_linux_sandbox::run_main()
}

View File

@@ -0,0 +1,3 @@
// Single integration test binary that aggregates all test modules.
// The submodules live in `tests/suite/`.
mod suite;

View File

@@ -0,0 +1,238 @@
#![cfg(target_os = "linux")]
use codex_core::config::types::ShellEnvironmentPolicy;
use codex_core::error::CodexErr;
use codex_core::error::SandboxErr;
use codex_core::exec::ExecParams;
use codex_core::exec::SandboxType;
use codex_core::exec::process_exec_tool_call;
use codex_core::exec_env::create_env;
use codex_core::protocol::SandboxPolicy;
use std::collections::HashMap;
use std::path::PathBuf;
use tempfile::NamedTempFile;
// At least on GitHub CI, the arm64 tests appear to need longer timeouts.
#[cfg(not(target_arch = "aarch64"))]
const SHORT_TIMEOUT_MS: u64 = 200;
#[cfg(target_arch = "aarch64")]
const SHORT_TIMEOUT_MS: u64 = 5_000;
#[cfg(not(target_arch = "aarch64"))]
const LONG_TIMEOUT_MS: u64 = 1_000;
#[cfg(target_arch = "aarch64")]
const LONG_TIMEOUT_MS: u64 = 5_000;
#[cfg(not(target_arch = "aarch64"))]
const NETWORK_TIMEOUT_MS: u64 = 2_000;
#[cfg(target_arch = "aarch64")]
const NETWORK_TIMEOUT_MS: u64 = 10_000;
fn create_env_from_core_vars() -> HashMap<String, String> {
let policy = ShellEnvironmentPolicy::default();
create_env(&policy)
}
#[expect(clippy::print_stdout, clippy::expect_used, clippy::unwrap_used)]
async fn run_cmd(cmd: &[&str], writable_roots: &[PathBuf], timeout_ms: u64) {
let cwd = std::env::current_dir().expect("cwd should exist");
let sandbox_cwd = cwd.clone();
let params = ExecParams {
command: cmd.iter().copied().map(str::to_owned).collect(),
cwd,
timeout_ms: Some(timeout_ms),
env: create_env_from_core_vars(),
with_escalated_permissions: None,
justification: None,
arg0: None,
};
let sandbox_policy = SandboxPolicy::WorkspaceWrite {
writable_roots: writable_roots.to_vec(),
network_access: false,
// Exclude tmp-related folders from writable roots because we need a
// folder that is writable by tests but that we intentionally disallow
// writing to in the sandbox.
exclude_tmpdir_env_var: true,
exclude_slash_tmp: true,
};
let sandbox_program = env!("CARGO_BIN_EXE_codex-linux-sandbox");
let codex_linux_sandbox_exe = Some(PathBuf::from(sandbox_program));
let res = process_exec_tool_call(
params,
SandboxType::LinuxSeccomp,
&sandbox_policy,
sandbox_cwd.as_path(),
&codex_linux_sandbox_exe,
None,
)
.await
.unwrap();
if res.exit_code != 0 {
println!("stdout:\n{}", res.stdout.text);
println!("stderr:\n{}", res.stderr.text);
panic!("exit code: {}", res.exit_code);
}
}
#[tokio::test]
async fn test_root_read() {
run_cmd(&["ls", "-l", "/bin"], &[], SHORT_TIMEOUT_MS).await;
}
#[tokio::test]
#[should_panic]
async fn test_root_write() {
let tmpfile = NamedTempFile::new().unwrap();
let tmpfile_path = tmpfile.path().to_string_lossy();
run_cmd(
&["bash", "-lc", &format!("echo blah > {tmpfile_path}")],
&[],
SHORT_TIMEOUT_MS,
)
.await;
}
#[tokio::test]
async fn test_dev_null_write() {
run_cmd(
&["bash", "-lc", "echo blah > /dev/null"],
&[],
// We have seen timeouts when running this test in CI on GitHub,
// so we are using a generous timeout until we can diagnose further.
LONG_TIMEOUT_MS,
)
.await;
}
#[tokio::test]
async fn test_writable_root() {
let tmpdir = tempfile::tempdir().unwrap();
let file_path = tmpdir.path().join("test");
run_cmd(
&[
"bash",
"-lc",
&format!("echo blah > {}", file_path.to_string_lossy()),
],
&[tmpdir.path().to_path_buf()],
// We have seen timeouts when running this test in CI on GitHub,
// so we are using a generous timeout until we can diagnose further.
LONG_TIMEOUT_MS,
)
.await;
}
#[tokio::test]
#[should_panic(expected = "Sandbox(Timeout")]
async fn test_timeout() {
run_cmd(&["sleep", "2"], &[], 50).await;
}
/// Helper that runs `cmd` under the Linux sandbox and asserts that the command
/// does NOT succeed (i.e. returns a nonzero exit code) **unless** the binary
/// is missing in which case we silently treat it as an accepted skip so the
/// suite remains green on leaner CI images.
#[expect(clippy::expect_used)]
async fn assert_network_blocked(cmd: &[&str]) {
let cwd = std::env::current_dir().expect("cwd should exist");
let sandbox_cwd = cwd.clone();
let params = ExecParams {
command: cmd.iter().copied().map(str::to_owned).collect(),
cwd,
// Give the tool a generous 2-second timeout so even slow DNS timeouts
// do not stall the suite.
timeout_ms: Some(NETWORK_TIMEOUT_MS),
env: create_env_from_core_vars(),
with_escalated_permissions: None,
justification: None,
arg0: None,
};
let sandbox_policy = SandboxPolicy::new_read_only_policy();
let sandbox_program = env!("CARGO_BIN_EXE_codex-linux-sandbox");
let codex_linux_sandbox_exe: Option<PathBuf> = Some(PathBuf::from(sandbox_program));
let result = process_exec_tool_call(
params,
SandboxType::LinuxSeccomp,
&sandbox_policy,
sandbox_cwd.as_path(),
&codex_linux_sandbox_exe,
None,
)
.await;
let output = match result {
Ok(output) => output,
Err(CodexErr::Sandbox(SandboxErr::Denied { output })) => *output,
_ => {
panic!("expected sandbox denied error, got: {result:?}");
}
};
dbg!(&output.stderr.text);
dbg!(&output.stdout.text);
dbg!(&output.exit_code);
// A completely missing binary exits with 127. Anything else should also
// be nonzero (EPERM from seccomp will usually bubble up as 1, 2, 13…)
// If—*and only if*—the command exits 0 we consider the sandbox breached.
if output.exit_code == 0 {
panic!(
"Network sandbox FAILED - {cmd:?} exited 0\nstdout:\n{}\nstderr:\n{}",
output.stdout.text, output.stderr.text
);
}
}
#[tokio::test]
async fn sandbox_blocks_curl() {
assert_network_blocked(&["curl", "-I", "http://openai.com"]).await;
}
#[tokio::test]
async fn sandbox_blocks_wget() {
assert_network_blocked(&["wget", "-qO-", "http://openai.com"]).await;
}
#[tokio::test]
async fn sandbox_blocks_ping() {
// ICMP requires raw socket should be denied quickly with EPERM.
assert_network_blocked(&["ping", "-c", "1", "8.8.8.8"]).await;
}
#[tokio::test]
async fn sandbox_blocks_nc() {
// Zerolength connection attempt to localhost.
assert_network_blocked(&["nc", "-z", "127.0.0.1", "80"]).await;
}
#[tokio::test]
async fn sandbox_blocks_ssh() {
// Force ssh to attempt a real TCP connection but fail quickly. `BatchMode`
// avoids password prompts, and `ConnectTimeout` keeps the hang time low.
assert_network_blocked(&[
"ssh",
"-o",
"BatchMode=yes",
"-o",
"ConnectTimeout=1",
"github.com",
])
.await;
}
#[tokio::test]
async fn sandbox_blocks_getent() {
assert_network_blocked(&["getent", "ahosts", "openai.com"]).await;
}
#[tokio::test]
async fn sandbox_blocks_dev_tcp_redirection() {
// This syntax is only supported by bash and zsh. We try bash first.
// Fallback generic socket attempt using /bin/sh with bashstyle /dev/tcp. Not
// all images ship bash, so we guard against 127 as well.
assert_network_blocked(&["bash", "-c", "echo hi > /dev/tcp/127.0.0.1/80"]).await;
}

View File

@@ -0,0 +1,2 @@
// Aggregates all former standalone integration tests as modules.
mod landlock;