Files
agent-harness/agent-security/tools/sandbox.py
Roger Oriol 0bb3b1c601 cleanup
2026-07-21 19:17:36 +02:00

258 lines
9.4 KiB
Python

"""Docker-based sandbox for tool execution.
Instead of confining tools with in-process path checks and a command
denylist (which is only as strong as the checks we remember to write),
the action tools run inside a long-lived Docker container. The user's
project is bind-mounted into the container; everything outside that
mount is the container's own minimal filesystem and is invisible or
read-only to the tool. Network egress can be disabled entirely with
``--network none``.
The container is started once per session and reused for every action
tool call (via ``docker exec``) to avoid per-call startup latency.
In-memory planning tools (todo, scratchpad, ask_question) are **not**
run in the container — their state would not survive between separate
``docker exec`` processes — so they stay in-process on the host.
This module requires a Docker-compatible container CLI on the host:
Docker, or Podman (which is auto-detected when ``docker info`` does not
work — note that a shell ``alias docker=podman`` is **not** enough,
because the agent invokes the binary directly via ``subprocess`` without
a shell). The runtime can also be forced with the ``$AGENT_DOCKER``
environment variable. On first use the ``agent-security-runner`` image
is built automatically from the ``Dockerfile`` next to this package.
"""
import json
import os
import shlex
import subprocess
import uuid
from pathlib import Path
class DockerSandboxError(Exception):
"""Raised when the sandbox container cannot be used to run a tool."""
# Tools that touch the outside world and therefore run in the container.
ACTION_TOOLS = {
"read_file",
"glob_files",
"grep",
"write_file",
"edit_file",
"run_bash",
"webfetch",
}
DEFAULT_IMAGE = "agent-security-runner"
# Default per-tool-call timeout. 30 minutes was far too generous
# and let a hanging command block the whole session. Lowered to 120 s
# with a --tool-timeout CLI override.
EXEC_TIMEOUT_S = 120
def _runtime_works(cmd: list[str]) -> bool:
"""Return True if the given runtime CLI can talk to its daemon."""
try:
return subprocess.run(
cmd + ["info"], capture_output=True
).returncode == 0
except FileNotFoundError:
return False
def _resolve_runtime() -> list[str]:
"""Pick the container-runtime command to use.
Preference order:
1. ``$AGENT_DOCKER`` (explicit override, e.g. ``podman`` or
``/usr/bin/podman``; may include arguments such as ``sudo podman``);
2. ``docker`` if ``docker info`` succeeds;
3. ``podman`` if ``podman info`` succeeds;
4. ``docker`` as a last resort, so the caller raises the standard,
informative error instead of failing obscurely.
"""
override = os.environ.get("AGENT_DOCKER", "").strip()
candidates: list[list[str]] = []
if override:
candidates.append(shlex.split(override))
candidates.append(["docker"])
candidates.append(["podman"])
seen: set[tuple[str, ...]] = set()
for cmd in candidates:
key = tuple(cmd)
if key in seen:
continue
seen.add(key)
if _runtime_works(cmd):
return cmd
return ["docker"]
# Resolved once at import; re-checked per instance in case the daemon
# was stopped between import and sandbox creation.
DOCKER_CMD = _resolve_runtime()
class DockerSandbox:
"""Manage a long-lived container that executes action tool calls."""
def __init__(
self,
project_root: Path,
tools_dir: Path,
network: str = "bridge",
image: str = DEFAULT_IMAGE,
build_context: Path | None = None,
exec_timeout: float = EXEC_TIMEOUT_S,
container_env: dict | None = None,
):
self.runtime = DOCKER_CMD
if not _runtime_works(self.runtime):
raise DockerSandboxError(
"No container runtime is available on the host. Install "
"Docker or Podman and ensure its daemon/service is running, "
"or set $AGENT_DOCKER to the binary to use. (Note: a shell "
"alias such as `alias docker=podman` is not enough, because "
"the agent invokes the CLI directly without a shell.)"
)
self.project_root = Path(project_root).resolve()
if not self.project_root.is_dir():
raise DockerSandboxError(
f"Project root is not a directory: {self.project_root}"
)
self.tools_dir = Path(tools_dir).resolve()
self.build_context = Path(
build_context or self.tools_dir.parent).resolve()
self.image = image
self.network = network
self.exec_timeout = float(exec_timeout)
# the harness — not the model — controls the container env.
# Only an allowlist of vars is inherited from the host; secret-
# looking env vars are stripped before the container ever starts.
self.container_env = container_env or {}
self.container = f"agent-sandbox-{uuid.uuid4().hex[:8]}"
self._ensure_image()
self._start_container()
# -- image ----------------------------------------------------------
def _ensure_image(self) -> None:
inspect = subprocess.run(
self.runtime + ["image", "inspect", self.image],
capture_output=True,
)
if inspect.returncode == 0:
return
dockerfile = self.build_context / "Dockerfile"
if not dockerfile.exists():
raise DockerSandboxError(
f"Cannot build sandbox image: Dockerfile not found at {
dockerfile}."
)
print(f" [sandbox] building image '{self.image}' (one-time)...")
build = subprocess.run(
self.runtime + ["build", "-t",
self.image, str(self.build_context)],
)
if build.returncode != 0:
raise DockerSandboxError(
f"Failed to build sandbox image '{self.image}' "
f"(docker build exited {build.returncode})."
)
# -- container lifecycle -------------------------------------------
def _start_container(self) -> None:
uid = os.getuid() if hasattr(os, "getuid") else 0
gid = os.getgid() if hasattr(os, "getgid") else 0
cmd = [
*self.runtime, "run", "-d",
"--name", self.container,
"--network", self.network,
"--user", f"{uid}:{gid}",
# Mount the project at the same absolute path so paths the
# agent reports match between host and container.
"-v", f"{self.project_root}:{self.project_root}",
# Mount the tool implementations read-only.
"-v", f"{self.tools_dir}:/agent_tools:ro",
"-w", str(self.project_root),
]
# Credential injection at the harness level: only the
# allowlisted env vars (set by the harness, never by the model)
# are passed to the container. Host credentials are stripped.
for name, value in self.container_env.items():
cmd.extend(["-e", f"{name}={value}"])
cmd.extend([
"--rm",
self.image,
"sleep", "infinity",
])
run = subprocess.run(cmd, capture_output=True, text=True)
if run.returncode != 0:
raise DockerSandboxError(
f"Could not start sandbox container: {
run.stderr.strip() or run.stdout.strip()}"
)
# Sanity check: confirm the container is actually running.
ps = subprocess.run(
self.runtime + ["inspect", "-f",
"{{.State.Running}}", self.container],
capture_output=True, text=True,
)
if ps.returncode != 0 or ps.stdout.strip() != "true":
raise DockerSandboxError(
f"Sandbox container '{
self.container}' is not running after start."
)
# -- tool execution ------------------------------------------------
def run_tool(self, name: str, args: dict) -> str:
"""Execute *name* with *args* inside the container, return its output.
§4.3: enforces a per-call timeout (``self.exec_timeout``). A
timeout is reported back as a ``DockerSandboxError`` with a
clear message so the LLM knows not to retry blindly.
"""
try:
proc = subprocess.run(
[
*self.runtime, "exec", "-i",
self.container,
"python", "/agent_tools/_dispatch.py", name,
],
input=json.dumps(args),
capture_output=True,
text=True,
timeout=self.exec_timeout,
)
except subprocess.TimeoutExpired:
raise DockerSandboxError(
f"Tool '{name}' timed out after {self.exec_timeout:.0f}s. "
"The command did not finish in the allowed time. Do not "
"retry the same call — adjust the approach or ask the user."
)
if proc.returncode != 0:
err = (proc.stderr or proc.stdout or "").strip()
raise DockerSandboxError(
f"Container exec for '{name}' failed (exit {proc.returncode}): {
err}"
)
return proc.stdout
def close(self) -> None:
subprocess.run(
self.runtime + ["rm", "-f", self.container], capture_output=True
)