Compare commits

...
27 changed files with 866 additions and 738 deletions
+112
View File
@@ -0,0 +1,112 @@
name: Sandbox Image
on:
workflow_dispatch:
inputs:
tag:
description: 'Image tag to publish (e.g. 1.2.0)'
required: true
latest:
description: 'Also tag as latest'
type: boolean
default: true
permissions:
contents: read
env:
IMAGE: ghcr.io/${{ github.repository_owner }}/strix-sandbox
jobs:
build:
strategy:
fail-fast: false
matrix:
include:
- os: ubuntu-22.04
platform: linux/amd64
- os: ubuntu-22.04-arm
platform: linux/arm64
runs-on: ${{ matrix.os }}
permissions:
contents: read
packages: write
steps:
- uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
with:
persist-credentials: false
- uses: docker/setup-buildx-action@bb05f3f5519dd87d3ba754cc423b652a5edd6d2c # v4.2.0
- uses: docker/login-action@dbcb813823bdd20940b903addbd779551569679f # v4.6.0
with:
registry: ghcr.io
username: ${{ github.actor }}
password: ${{ secrets.GITHUB_TOKEN }}
- name: Build and push by digest
id: build
uses: docker/build-push-action@53b7df96c91f9c12dcc8a07bcb9ccacbed38856a # v7.3.0
with:
context: .
file: containers/Dockerfile
platforms: ${{ matrix.platform }}
provenance: mode=max
sbom: true
outputs: type=image,name=${{ env.IMAGE }},push-by-digest=true,name-canonical=true,push=true
- name: Export digest
env:
DIGEST: ${{ steps.build.outputs.digest }}
run: |
set -euo pipefail
mkdir -p /tmp/digests
touch "/tmp/digests/${DIGEST#sha256:}"
- uses: actions/upload-artifact@043fb46d1a93c77aae656e7c1c64a875d1fc6a0a # v7.0.1
with:
name: digest-${{ runner.arch }}
path: /tmp/digests/*
if-no-files-found: error
retention-days: 1
publish:
needs: build
runs-on: ubuntu-latest
permissions:
contents: read
packages: write
steps:
- uses: actions/download-artifact@3e5f45b2cfb9172054b4087a40e8e0b5a5461e7c # v8.0.1
with:
path: /tmp/digests
pattern: digest-*
merge-multiple: true
- uses: docker/setup-buildx-action@bb05f3f5519dd87d3ba754cc423b652a5edd6d2c # v4.2.0
- uses: docker/login-action@dbcb813823bdd20940b903addbd779551569679f # v4.6.0
with:
registry: ghcr.io
username: ${{ github.actor }}
password: ${{ secrets.GITHUB_TOKEN }}
- name: Create manifest list
env:
TAG: ${{ inputs.tag }}
ALSO_LATEST: ${{ inputs.latest }}
run: |
set -euo pipefail
tags=(-t "${IMAGE}:${TAG}")
if [ "${ALSO_LATEST}" = "true" ]; then
tags+=(-t "${IMAGE}:latest")
fi
digests=()
for file in /tmp/digests/*; do
digests+=("${IMAGE}@sha256:$(basename "$file")")
done
docker buildx imagetools create "${tags[@]}" "${digests[@]}"
docker buildx imagetools inspect "${IMAGE}:${TAG}"
+2 -2
View File
@@ -162,6 +162,7 @@ USER root
ARG TRUFFLEHOG_VERSION=3.95.9
RUN curl -sSfL https://raw.githubusercontent.com/trufflesecurity/trufflehog/main/scripts/install.sh | sh -s -- -b /home/pentester/.local/bin "v${TRUFFLEHOG_VERSION}" && \
chown -R pentester:pentester /home/pentester/.local
ARG GITLEAKS_VERSION=8.30.1
RUN set -eux; \
ARCH="$(uname -m)"; \
case "$ARCH" in \
@@ -169,8 +170,7 @@ RUN set -eux; \
aarch64|arm64) GITLEAKS_ARCH="arm64" ;; \
*) echo "Unsupported architecture: $ARCH" >&2; exit 1 ;; \
esac; \
TAG="$(curl -fsSL https://api.github.com/repos/gitleaks/gitleaks/releases/latest | jq -r .tag_name)"; \
curl -fsSL "https://github.com/gitleaks/gitleaks/releases/download/${TAG}/gitleaks_${TAG#v}_linux_${GITLEAKS_ARCH}.tar.gz" -o /tmp/gitleaks.tgz; \
curl -fsSL "https://github.com/gitleaks/gitleaks/releases/download/v${GITLEAKS_VERSION}/gitleaks_${GITLEAKS_VERSION}_linux_${GITLEAKS_ARCH}.tar.gz" -o /tmp/gitleaks.tgz; \
tar -xzf /tmp/gitleaks.tgz -C /tmp; \
install -m 0755 /tmp/gitleaks /usr/local/bin/gitleaks; \
rm -f /tmp/gitleaks /tmp/gitleaks.tgz
+16
View File
@@ -1,6 +1,22 @@
#!/bin/bash
set -e
if [ -n "${STRIX_HOST_UID:-}" ] && [ "${STRIX_HOST_UID}" != "0" ] && [ "${STRIX_HOST_UID}" != "$(id -u)" ]; then
exec sudo -E -- bash -c '
set -e
gid="${STRIX_HOST_GID:-$STRIX_HOST_UID}"
old_uid="$1"
old_gid="$2"
export PATH="$3"
shift 3
sed -i "s|^pentester:x:${old_uid}:${old_gid}:|pentester:x:${STRIX_HOST_UID}:${gid}:|" /etc/passwd
sed -i "s|^pentester:x:${old_gid}:|pentester:x:${gid}:|" /etc/group
chown -R "${STRIX_HOST_UID}:${gid}" /home/pentester /app/certs
chown "${STRIX_HOST_UID}:${gid}" /workspace
exec setpriv --reuid "${STRIX_HOST_UID}" --regid "${gid}" --init-groups "$0" "$@"
' "$0" "$(id -u)" "$(id -g)" "$PATH" "$@"
fi
CAIDO_PORT=48080
CAIDO_LOG="/tmp/caido_startup.log"
+1 -5
View File
@@ -106,7 +106,7 @@ When remote vars are set, Strix dual-writes telemetry to both local JSONL and th
## Docker Configuration
<ParamField path="STRIX_IMAGE" default="ghcr.io/usestrix/strix-sandbox:1.0.0" type="string">
<ParamField path="STRIX_IMAGE" default="ghcr.io/usestrix/strix-sandbox:1.2.0" type="string">
Docker image to use for the sandbox container.
</ParamField>
@@ -118,10 +118,6 @@ When remote vars are set, Strix dual-writes telemetry to both local JSONL and th
Runtime backend for the sandbox environment.
</ParamField>
<ParamField path="STRIX_MAX_LOCAL_COPY_MB" default="1024" type="integer">
Maximum size (in MB) of a local directory target that Strix will copy into the sandbox file-by-file. Larger targets exit early with a suggestion to use `--mount` instead. Set to `0` to disable the check.
</ParamField>
## Sandbox Configuration
<ParamField path="STRIX_SANDBOX_EXECUTION_TIMEOUT" default="120" type="integer">
+6 -19
View File
@@ -6,33 +6,23 @@ description: "Command-line options for Strix"
## Basic Usage
```bash
strix (--target <target> | --target-list <path> | --mount <path>) [options]
strix (--target <target> | --target-list <path>) [options]
```
## Options
<ParamField path="--target, -t" type="string">
Target to test. Accepts URLs, repositories, local directories, domains, or IP addresses. Can be specified multiple times. Fresh runs require at least one target source: `--target`, `--target-list`, or `--mount`.
Target to test. Accepts URLs, repositories, local directories, domains, or IP addresses. Can be specified multiple times. Fresh runs require at least one target source: `--target` or `--target-list`.
<Note>
A local directory is mounted into the sandbox live and **writable**, so the agent edits your real files (`.git` excepted). Commit or stash first.
</Note>
</ParamField>
<ParamField path="--target-list" type="string">
Path to a file containing targets, one per non-empty, non-comment line. Lines starting with `#` are ignored. Can be specified multiple times and combined with `--target`.
</ParamField>
<ParamField path="--mount" type="string">
Bind-mount a local directory into the sandbox (read-only) instead of copying it in file-by-file. Use this for large repositories that are too big to stream into the container. Can be specified multiple times.
Strix copies local `--target` directories into the sandbox one file at a time, which stalls on very large trees. When a local target exceeds the copy limit (see `STRIX_MAX_LOCAL_COPY_MB`, default 1024 MB) Strix exits early and asks you to re-run with `--mount`.
<Note>
The mount is read-only to protect your source from accidental modification. This is not a hard security boundary: a root process inside the container can remount it writable, so treat `--mount` as "scan my own code", not as isolation from untrusted code.
</Note>
<Note>
The size pre-flight only covers local directory targets. Remote repositories (cloned at scan time) are not size-checked.
</Note>
</ParamField>
<ParamField path="--instruction" type="string">
Custom instructions for the scan. Use for credentials, focus areas, or specific testing approaches.
</ParamField>
@@ -140,9 +130,6 @@ strix -t https://github.com/org/app -t https://staging.example.com
# Targets from a file
strix --target-list ./targets.txt
# Large local repository — bind-mount instead of copying it in
strix --mount ./huge-monorepo
```
## Exit Codes
+1 -1
View File
@@ -4,7 +4,7 @@ set -euo pipefail
APP=strix
REPO="usestrix/strix"
STRIX_IMAGE="ghcr.io/usestrix/strix-sandbox:1.1.0"
STRIX_IMAGE="ghcr.io/usestrix/strix-sandbox:1.2.0"
MUTED='\033[0;2m'
RED='\033[0;31m'
+1 -6
View File
@@ -97,15 +97,10 @@ class RuntimeSettings(BaseSettings):
model_config = _BASE_CONFIG
image: str = Field(
default="ghcr.io/usestrix/strix-sandbox:1.1.0",
default="ghcr.io/usestrix/strix-sandbox:1.2.0",
alias="STRIX_IMAGE",
)
backend: str = Field(default="docker", alias="STRIX_RUNTIME_BACKEND")
# Hard cap on a local target's size before we refuse to stream it into the
# sandbox file-by-file (the SDK copies every file individually, which stalls
# on large repos). Above this, the user must bind-mount via ``--mount``.
# Set to 0 (or less) to disable the pre-flight check entirely.
max_local_copy_mb: int = Field(default=1024, alias="STRIX_MAX_LOCAL_COPY_MB")
# Max screenshot/image tool outputs kept live per agent context (0 = none).
max_context_images: int = Field(default=3, ge=0, alias="STRIX_MAX_CONTEXT_IMAGES")
+22 -2
View File
@@ -55,6 +55,7 @@ class AgentCoordinator:
self.idle_resume_counts: dict[str, int] = {}
self.wait_kinds: dict[str, WaitKind] = {}
self.runtimes: dict[str, AgentRuntime] = {}
self._parent_notified: set[str] = set()
self._lock = asyncio.Lock()
self._snapshot_path: Path | None = None
self.is_shutting_down = False
@@ -191,6 +192,7 @@ class AgentCoordinator:
self.errors.pop(agent_id, None)
self.wait_kinds.pop(agent_id, None)
self.runtimes.setdefault(agent_id, AgentRuntime()).user_wake_required = False
self._parent_notified.discard(agent_id)
await self._maybe_snapshot()
async def park_waiting(self, agent_id: str, *, wait_kind: WaitKind) -> None:
@@ -252,12 +254,27 @@ class AgentCoordinator:
self.errors[agent_id] = error
elif status == "running":
self.errors.pop(agent_id, None)
if status == "running":
# Running again means a fresh stint that owes its parent its own notice.
self._parent_notified.discard(agent_id)
runtime = self.runtimes.setdefault(agent_id, AgentRuntime())
runtime.user_wake_required = status in {"failed", "crashed"}
runtime.wake.set()
logger.info("agent.status %s=%s", agent_id, status)
await self._maybe_snapshot()
async def claim_parent_notice(self, agent_id: str) -> bool:
"""Reserve the one notice a child owes its parent when it stops running.
A completion report and a terminal notice carry the same information, so
whichever comes first claims the slot and the other is skipped.
"""
async with self._lock:
if agent_id in self._parent_notified:
return False
self._parent_notified.add(agent_id)
return True
async def send(
self, target_agent_id: str, message: dict[str, Any], *, interrupt: bool = True
) -> bool:
@@ -366,12 +383,15 @@ class AgentCoordinator:
if tasks:
await asyncio.gather(*tasks, return_exceptions=True)
async def cancel_descendants_graceful(self, agent_id: str) -> None:
async def cancel_descendants_graceful(self, agent_id: str) -> list[str]:
"""Stop a subtree leaves-first and report which agents were stopped."""
async with self._lock:
order = self._subtree_order_locked(agent_id)
for aid in reversed(order):
stopped = list(reversed(order))
for aid in stopped:
await self.request_stop(aid)
await self._maybe_snapshot()
return stopped
async def attach_stream(
self,
+33 -8
View File
@@ -539,7 +539,7 @@ async def _exhausted_recovery(
"""
if not interactive:
await coordinator.set_status(agent_id, "crashed")
await _notify_parent_on_terminal(coordinator, agent_id, "crashed")
await notify_parent_on_terminal(coordinator, agent_id, "crashed")
raise MaxTurnsExceeded(
"Agent exhausted recovery attempts without calling finish_scan or agent_finish."
)
@@ -622,7 +622,7 @@ async def _run_cycle_parked(
except Exception as exc:
logger.exception("error escaped the run cycle for %s; parking as failed", agent_id)
await coordinator.set_status(agent_id, "failed", error=str(exc) or type(exc).__name__)
await _notify_parent_on_terminal(coordinator, agent_id, "failed")
await notify_parent_on_terminal(coordinator, agent_id, "failed")
return None
@@ -778,7 +778,7 @@ async def _run_cycle( # noqa: PLR0912, PLR0915
if isinstance(exc, ProviderRefusalError):
logger.warning("agent %s refused by the model provider: %s", agent_id, exc)
await coordinator.set_status(agent_id, "failed", error=str(exc))
await _notify_parent_on_terminal(coordinator, agent_id, "failed")
await notify_parent_on_terminal(coordinator, agent_id, "failed")
return None
if not interactive:
raise
@@ -790,7 +790,7 @@ async def _run_cycle( # noqa: PLR0912, PLR0915
status = "crashed"
logger.exception("agent run failed for %s; parking as %s", agent_id, status)
await coordinator.set_status(agent_id, status, error=str(exc) or type(exc).__name__)
await _notify_parent_on_terminal(coordinator, agent_id, status)
await notify_parent_on_terminal(coordinator, agent_id, status)
return None
else:
return cast("RunResultBase | None", stream)
@@ -851,6 +851,11 @@ async def _append_tool_required_message(
_TERMINAL_NOTICE = {
"completed": (
"[Agent completed] {name} ({agent_id}) finished and is no longer running, but it "
"sent no completion report. Stop waiting on this child; ask it directly if you "
"need its results."
),
"crashed": (
"[Agent crash] {name} ({agent_id}) terminated unexpectedly. "
"Stop waiting on this child unless you want to message it again."
@@ -861,9 +866,9 @@ _TERMINAL_NOTICE = {
"message it again."
),
"stopped": (
"[Agent capped] {name} ({agent_id}) hit its turn limit and was stopped "
"before finishing. It will not send a completion report, so stop waiting "
"on this child; account for its capped subtask and continue."
"[Agent stopped] {name} ({agent_id}) was stopped before finishing (turn limit "
"or an explicit stop). It will not send a completion report, so stop waiting "
"on this child; account for its unfinished subtask and continue."
),
}
@@ -898,7 +903,7 @@ async def _notify_parent_on_stall(
)
async def _notify_parent_on_terminal(
async def notify_parent_on_terminal(
coordinator: AgentCoordinator,
agent_id: str,
status: str,
@@ -911,6 +916,8 @@ async def _notify_parent_on_terminal(
name = coordinator.names.get(agent_id, agent_id)
if parent is None:
return
if not await coordinator.claim_parent_notice(agent_id):
return
await coordinator.send(
parent,
{
@@ -945,6 +952,21 @@ async def _notify_root_on_budget_reserve(coordinator: AgentCoordinator) -> None:
await coordinator.send(root, _reserve_notice())
async def _notify_parent_on_exit(
coordinator: AgentCoordinator,
agent_id: str,
) -> None:
"""Backstop for a child whose loop ended without telling its parent.
Every terminal state counts, including ``completed``: a child that skips its
completion report leaves the parent waiting on a message nobody will send.
"""
status = await _agent_status(coordinator, agent_id)
if status is None:
return
await notify_parent_on_terminal(coordinator, agent_id, status)
async def _start_child_runner(
*,
parent_ctx: dict[str, Any],
@@ -998,6 +1020,9 @@ async def _start_child_runner(
logger.info("child %s stopped after reaching the scan budget limit", child_id)
except SubagentBudgetReservedError:
logger.info("child %s stopped at the sub-agent budget reserve", child_id)
finally:
if not coordinator.is_shutting_down:
await _notify_parent_on_exit(coordinator, child_id)
task_handle = asyncio.create_task(_child_loop(), name=f"agent-{name}-{child_id}")
await coordinator.attach_runtime(child_id, task=task_handle)
+5 -2
View File
@@ -59,8 +59,11 @@ def build_root_task(scan_config: dict[str, Any]) -> str:
)
elif ttype == "local_code":
path = details.get("target_path", "unknown")
suffix = ", read-only mount" if details.get("mount") else ""
sections["Local Codebases"].append(f"- {path} (available at: {workspace_path}{suffix})")
sections["Local Codebases"].append(
f"- {path} (available at: {workspace_path}; "
"this is the user's real directory, mounted live and writable — "
".git/.agents/.codex are read-only)"
)
elif ttype == "web_application":
sections["URLs"].append(f"- {details.get('target_url', '')}")
elif ttype == "ip_address":
+10
View File
@@ -53,6 +53,8 @@ if TYPE_CHECKING:
from agents.memory import SQLiteSession
from agents.result import RunResultBase
from strix.runtime.status import StatusSink
logger = logging.getLogger(__name__)
@@ -120,6 +122,7 @@ async def run_strix_scan(
event_sink: StreamEventSink | None = None,
root_instructions_override: str | None = None,
extra_system_prompt_context: dict[str, Any] | None = None,
status_sink: StatusSink | None = None,
) -> RunResultBase | None:
"""Run or resume one Strix scan against a sandbox.
@@ -129,6 +132,11 @@ async def run_strix_scan(
context before prompt rendering. Child agents keep the standard scan prompt
and context.
"""
def report(phase: str) -> None:
if status_sink is not None:
status_sink(phase)
if scan_id is None:
scan_id = f"scan-{uuid.uuid4().hex[:8]}"
@@ -219,7 +227,9 @@ async def run_strix_scan(
scan_id,
image=image,
local_sources=local_sources or [],
status_sink=status_sink,
)
report("Waiting for the first model response")
logger.info("Sandbox ready for scan %s", scan_id)
sandbox_session = bundle["session"]
+11
View File
@@ -21,6 +21,7 @@ from strix.runtime import session_manager
from .utils import (
build_live_stats_text,
format_vulnerability_report,
has_model_response,
)
@@ -135,11 +136,17 @@ async def run_cli(args: Any) -> None: # noqa: PLR0915
set_global_report_state(report_state)
startup_phase: list[str] = ["Starting up"]
def create_live_status() -> Panel:
status_text = Text()
status_text.append("Penetration test in progress", style="bold #22c55e")
status_text.append("\n\n")
if not has_model_response(report_state):
status_text.append(f"{startup_phase[0]}...", style="dim")
status_text.append("\n\n")
stats_text = build_live_stats_text(report_state)
if stats_text:
status_text.append(stats_text)
@@ -152,6 +159,9 @@ async def run_cli(args: Any) -> None: # noqa: PLR0915
padding=(1, 2),
)
def _note_startup_phase(phase: str) -> None:
startup_phase[:] = [phase]
try:
console.print()
@@ -186,6 +196,7 @@ async def run_cli(args: Any) -> None: # noqa: PLR0915
interactive=bool(getattr(args, "interactive", False)),
max_budget_usd=getattr(args, "max_budget_usd", None),
max_turns=getattr(args, "max_turns", DEFAULT_MAX_TURNS),
status_sink=_note_startup_phase,
)
finally:
stop_updates.set()
+16 -42
View File
@@ -33,12 +33,11 @@ from strix.interface.update_check import (
from strix.interface.utils import (
assign_workspace_subdirs,
build_final_stats_text,
build_mount_targets_info,
check_docker_connection,
check_mountable_dir,
clone_repository,
collect_local_sources,
dedupe_local_targets,
find_oversized_local_targets,
generate_run_name,
image_exists,
infer_target_type,
@@ -524,9 +523,6 @@ Examples:
# Local code analysis
strix --target ./my-project
# Large local repository (bind-mounted read-only instead of copied)
strix --mount ./huge-monorepo
# Domain penetration test
strix --target example.com
@@ -570,8 +566,9 @@ Examples:
type=str,
action="append",
help="Target to test (URL, repository, local directory path, domain name, or IP address). "
"Local directories are mounted into the sandbox writable. "
"Can be specified multiple times for multi-target scans. "
"Fresh runs require at least one of --target, --target-list, or --mount.",
"Fresh runs require --target or --target-list.",
)
parser.add_argument(
"--target-list",
@@ -581,15 +578,6 @@ Examples:
help="Path to a file containing targets, one per non-empty, non-comment line. "
"Can be specified multiple times and combined with --target.",
)
parser.add_argument(
"--mount",
type=str,
action="append",
metavar="PATH",
help="Bind-mount a local directory into the sandbox (read-only) instead of "
"copying it file-by-file. Use this for large repositories that are too big to "
"stream into the container. Can be specified multiple times.",
)
parser.add_argument(
"--instruction",
type=str,
@@ -720,9 +708,9 @@ Examples:
args.user_explicit_instruction = args.instruction if args.resume else None
if args.resume:
if args.target or args.target_list or args.mount:
if args.target or args.target_list:
parser.error(
"Cannot combine --resume with --target/--target-list/--mount. "
"Cannot combine --resume with --target/--target-list. "
"--resume picks up where the prior run left off, including the "
"original target list."
)
@@ -736,9 +724,9 @@ Examples:
f"or remove --resume to start over with the same targets."
)
else:
if not args.target and not args.target_list and not args.mount:
if not args.target and not args.target_list:
parser.error(
"the following arguments are required: -t/--target, --target-list, or --mount "
"the following arguments are required: -t/--target or --target-list "
"(or use --resume <run_name> to continue a prior scan)"
)
args.targets_info = []
@@ -761,33 +749,14 @@ Examples:
args.targets_info.append(
{"type": target_type, "details": target_dict, "original": display_target}
)
except ValueError:
parser.error(f"Invalid target '{target}'")
try:
args.targets_info.extend(build_mount_targets_info(args.mount or []))
except ValueError as e:
parser.error(str(e))
except ValueError as e:
parser.error(f"Invalid target '{target}': {e}")
args.targets_info = dedupe_local_targets(args.targets_info)
assign_workspace_subdirs(args.targets_info)
rewrite_localhost_targets(args.targets_info, HOST_GATEWAY_HOSTNAME)
max_local_copy_mb = load_settings().runtime.max_local_copy_mb
max_copy_bytes = max_local_copy_mb * 1024 * 1024
oversized = find_oversized_local_targets(args.targets_info, max_copy_bytes)
if oversized:
details = "; ".join(
f"{path} ({size / (1024 * 1024):.0f} MB)" for path, size in oversized
)
parser.error(
f"Local target too large to stream into the sandbox: {details}. "
f"The limit is {max_local_copy_mb} MB "
"(set STRIX_MAX_LOCAL_COPY_MB to change it). Re-run with "
"--mount <path> to bind-mount the directory instead of copying it."
)
return args
@@ -839,6 +808,12 @@ def _load_resume_state(args: argparse.Namespace, parser: argparse.ArgumentParser
if not isinstance(target, dict):
continue
details = target.get("details") or {}
if target.get("type") == "local_code" and details.get("target_path"):
try:
check_mountable_dir(Path(details["target_path"]).expanduser())
except ValueError as exc:
parser.error(f"--resume {args.resume}: {exc}")
continue
if target.get("type") != "repository":
continue
cloned = details.get("cloned_repo_path")
@@ -853,8 +828,7 @@ def _load_resume_state(args: argparse.Namespace, parser: argparse.ArgumentParser
if args.instruction is None:
args.instruction = state.get("instruction")
if state.get("local_sources"):
args.local_sources = state.get("local_sources")
args.local_sources = collect_local_sources(args.targets_info)
if state.get("diff_scope"):
args.diff_scope = state.get("diff_scope")
persisted_scan_mode = state.get("scan_mode")
+18 -1
View File
@@ -815,6 +815,8 @@ class StrixTUIApp(App): # type: ignore[misc]
self._scan_stop_event = threading.Event()
self._scan_completed = threading.Event()
self._scan_error: BaseException | None = None
self._startup_status = "Starting up"
self._startup_status_step = 0
self._error_noted_agents: set[str] = set()
self._budget_pause_notified = False
@@ -1111,7 +1113,9 @@ class StrixTUIApp(App): # type: ignore[misc]
self,
) -> tuple[Any, str | None]:
if not self.selected_agent_id:
return self._get_chat_placeholder_content("Loading...", "placeholder-no-agent")
return self._get_chat_placeholder_content(
f"{self._startup_status}...", f"placeholder-no-agent-{self._startup_status_step}"
)
events = self._gather_agent_events(self.selected_agent_id)
@@ -1525,6 +1529,7 @@ class StrixTUIApp(App): # type: ignore[misc]
max_budget_usd=getattr(self.args, "max_budget_usd", None),
max_turns=getattr(self.args, "max_turns", DEFAULT_MAX_TURNS),
event_sink=self._capture_sdk_event,
status_sink=self._capture_startup_status,
),
)
@@ -1556,6 +1561,18 @@ class StrixTUIApp(App): # type: ignore[misc]
self._scan_thread = threading.Thread(target=scan_target, daemon=True)
self._scan_thread.start()
def _capture_startup_status(self, phase: str) -> None:
try:
self.call_from_thread(self._record_startup_status, phase)
except RuntimeError:
self._record_startup_status(phase)
def _record_startup_status(self, phase: str) -> None:
self._startup_status = phase
self._startup_status_step += 1
if not self.show_splash and not self.selected_agent_id:
self.call_later(self._update_chat_view)
def _capture_sdk_event(self, agent_id: str, event: Any) -> None:
try:
self.call_from_thread(self._record_sdk_event, agent_id, event)
+101 -92
View File
@@ -290,6 +290,11 @@ def _detail_value(usage: dict[str, Any], detail_key: str, value_key: str) -> int
return _int_stat(details, value_key)
def has_model_response(report_state: Any) -> bool:
usage = _llm_usage(report_state)
return bool(usage) and _int_stat(usage, "requests") > 0
def _build_llm_usage_stats(
stats_text: Text,
report_state: Any,
@@ -1131,6 +1136,7 @@ def infer_target_type(target: str) -> tuple[str, dict[str, str]]: # noqa: PLR09
try:
if path.exists():
if path.is_dir():
check_mountable_dir(path)
return "local_code", {"target_path": str(path.resolve())}
raise ValueError(f"Path exists but is not a directory: {target}")
except (OSError, RuntimeError) as e:
@@ -1259,7 +1265,7 @@ def collect_local_sources(targets_info: list[dict[str, Any]]) -> list[dict[str,
{
"source_path": details["target_path"],
"workspace_subdir": workspace_subdir,
"mount": bool(details.get("mount", False)),
"protect_metadata": True,
}
)
@@ -1268,123 +1274,126 @@ def collect_local_sources(targets_info: list[dict[str, Any]]) -> list[dict[str,
{
"source_path": details["cloned_repo_path"],
"workspace_subdir": workspace_subdir,
"mount": False,
"protect_metadata": False,
}
)
return local_sources
def directory_size_bytes(path: Path) -> int:
"""Total size in bytes of regular files under ``path`` (symlinks not followed).
# Refused along with everything under them.
_FORBIDDEN_MOUNT_TREES = frozenset(
{
"/bin",
"/sbin",
"/usr",
"/etc",
"/lib",
"/lib64",
"/nix/store",
"/run/current-system/sw",
"/Applications",
"/Library",
"/System",
"/dev",
"/boot",
"/proc",
"/sys",
}
)
Best-effort: files that disappear or can't be stat'd mid-walk are skipped.
Used as a cheap (stat-only) pre-flight to estimate the cost of streaming a
local target into the sandbox before we actually try to copy it.
# Refused themselves, but they hold projects too, so their contents are fine.
_FORBIDDEN_MOUNT_ROOTS = frozenset(
{
"/",
"/private",
"/var",
"/opt",
"/home",
"/root",
"/srv",
"/Users",
"/Volumes",
}
)
Directories that can't be listed (e.g. permission denied) are logged and
skipped rather than silently dropped so an under-count is at least
visible but the returned total then excludes their contents.
"""
_FORBIDDEN_WINDOWS_TREE_NAMES = frozenset(
{"windows", "program files", "program files (x86)", "programdata"}
)
def _on_walk_error(error: OSError) -> None:
logger.warning("Could not read %s while measuring size: %s", error.filename, error)
total = 0
for root, _dirs, files in os.walk(path, followlinks=False, onerror=_on_walk_error):
for name in files:
file_path = os.path.join(root, name) # noqa: PTH118
try:
if os.path.islink(file_path): # noqa: PTH114
continue
total += os.path.getsize(file_path) # noqa: PTH202
except OSError:
continue
return total
_FORBIDDEN_MOUNT_DIR_NAMES = frozenset(
{
".ssh",
".tsh",
".brev",
".gnupg",
".aws",
".azure",
".kube",
".docker",
".config",
".npm",
".pki",
".terraform.d",
}
)
def find_oversized_local_targets(
targets_info: list[dict[str, Any]], max_bytes: int
) -> list[tuple[str, int]]:
"""Return ``(path, size_bytes)`` for non-mounted local targets over ``max_bytes``.
Mounted targets are bind-mounted rather than copied, so their size is
irrelevant and they are excluded. A ``max_bytes`` of zero or less disables
the check entirely (returns no targets).
"""
if max_bytes <= 0:
return []
oversized: list[tuple[str, int]] = []
for target in targets_info:
if target.get("type") != "local_code":
continue
details = target.get("details") or {}
if details.get("mount"):
continue
target_path = details.get("target_path")
if not target_path:
continue
size = directory_size_bytes(Path(target_path))
if size > max_bytes:
oversized.append((target_path, size))
return oversized
def _is_within(path: Path, ancestor: Path) -> bool:
ancestor_parts = [part.casefold() for part in ancestor.parts]
path_parts = [part.casefold() for part in path.parts]
return path_parts[: len(ancestor_parts)] == ancestor_parts
def build_mount_targets_info(mount_paths: list[str]) -> list[dict[str, Any]]:
"""Build ``targets_info`` entries for ``--mount`` directories.
def check_mountable_dir(path: Path) -> None:
resolved = path.resolve()
if not resolved.is_dir():
raise ValueError(f"'{path}' is not an existing directory.")
Each path must be an existing local directory; it is bind-mounted into the
sandbox (read-only) instead of being copied file-by-file. Raises
``ValueError`` for an empty path, or one that does not exist or is not a
directory.
"""
targets_info: list[dict[str, Any]] = []
for raw in mount_paths:
if not raw or not raw.strip():
raise ValueError("--mount path must not be empty.")
path = Path(raw).expanduser()
try:
resolved = path.resolve()
is_dir = resolved.is_dir()
except (OSError, RuntimeError) as e:
raise ValueError(f"Invalid mount path '{raw}': {e!s}") from e
if not is_dir:
raise ValueError(
f"Mount path '{raw}' is not an existing directory. "
"--mount requires a path to a local directory."
)
targets_info.append(
{
"type": "local_code",
"details": {"target_path": str(resolved), "mount": True},
"original": str(resolved),
}
# Both the literal and the resolved form: macOS reaches /etc through the
# /private/etc symlink, and only the resolved path is compared below.
exact = {str(Path(root)).casefold() for root in _FORBIDDEN_MOUNT_ROOTS}
exact |= {str(Path(root).resolve()).casefold() for root in _FORBIDDEN_MOUNT_ROOTS}
exact.add(str(Path.home().resolve()).casefold())
tree_roots = set(_FORBIDDEN_MOUNT_TREES)
if os.name == "nt":
drive = Path(resolved.anchor)
tree_roots |= {str(drive / name) for name in _FORBIDDEN_WINDOWS_TREE_NAMES}
exact.add(str(drive / "Users").casefold())
trees = [Path(root) for root in tree_roots] + [Path(root).resolve() for root in tree_roots]
if (
str(resolved).casefold() in exact
or resolved.parent == resolved
or any(_is_within(resolved, tree) for tree in trees)
):
raise ValueError(
f"Refusing to mount '{resolved}' into the sandbox: it is a system "
"or home directory, not a codebase. Point the target at the "
"project directory you want tested."
)
credential = next(
(part for part in resolved.parts if part.casefold() in _FORBIDDEN_MOUNT_DIR_NAMES), None
)
if credential is not None:
raise ValueError(
f"Refusing to mount '{resolved}' into the sandbox: '{credential}' "
"holds credentials, not code."
)
return targets_info
def dedupe_local_targets(targets_info: list[dict[str, Any]]) -> list[dict[str, Any]]:
"""Collapse local_code targets that resolve to the same path.
When a directory is supplied both as a copied ``--target`` and via
``--mount`` (or as duplicate values of either), keep one entry and prefer
the bind-mounted one so the same tree is never both streamed in and
mounted. Order is preserved; non-local targets pass through untouched.
"""
result: list[dict[str, Any]] = []
index_by_path: dict[str, int] = {}
seen_paths: set[str] = set()
for target in targets_info:
details = target.get("details") or {}
path = details.get("target_path")
if target.get("type") != "local_code" or not path:
result.append(target)
continue
existing = index_by_path.get(path)
if existing is None:
index_by_path[path] = len(result)
if path not in seen_paths:
seen_paths.add(path)
result.append(target)
elif details.get("mount") and not (result[existing].get("details") or {}).get("mount"):
result[existing] = target # bind mount supersedes the copied entry
return result
+25 -13
View File
@@ -31,16 +31,11 @@ async def _docker_backend(
``docker`` lazily so deployments that target a non-Docker
backend don't need the docker-py library installed.
``session.start()`` is what materializes the manifest entries
(LocalDir copies and manifest-declared volume/FUSE mounts) into the
running container the SDK's ``client.create()`` only builds the inner
session object without applying the manifest. ``async with session:``
would call it too, but Strix manages session lifetime explicitly via
``client.delete()`` so we trigger ``start()`` ourselves.
``bind_mounts`` are host directories (e.g. large repos passed via
``--mount``) bind-mounted read-only; unlike manifest entries they are
applied by Docker at container-create time, not by ``start()``.
``session.start()`` is what materializes the manifest into the running
container the SDK's ``client.create()`` only builds the inner session
object without applying it. ``async with session:`` would call it too, but
Strix manages session lifetime explicitly via ``client.delete()`` so we
trigger ``start()`` ourselves.
"""
import docker
from agents.sandbox.sandboxes.docker import DockerSandboxClientOptions
@@ -59,6 +54,8 @@ _BACKENDS: dict[str, SandboxBackend] = {
"docker": _docker_backend,
}
_BIND_MOUNT_BACKENDS: set[str] = {"docker"}
def get_backend(name: str) -> SandboxBackend:
"""Return the backend factory for ``name`` or raise.
@@ -78,15 +75,30 @@ def get_backend(name: str) -> SandboxBackend:
return backend
def register_backend(name: str, backend: SandboxBackend) -> None:
def register_backend(
name: str,
backend: SandboxBackend,
*,
supports_bind_mounts: bool = False,
) -> None:
"""Register a custom backend under ``name``.
Intended for downstream users who ship their own runtime register
before any ``session_manager.create_or_reuse`` call. Re-registering
an existing name overwrites the prior entry.
an existing name overwrites the prior entry. ``supports_bind_mounts``
defaults to False: a remote runtime cannot see the caller's filesystem, so
it is handed local sources as manifest entries to upload instead.
"""
_BACKENDS[name] = backend
logger.info("Registered sandbox backend: %s", name)
if supports_bind_mounts:
_BIND_MOUNT_BACKENDS.add(name)
else:
_BIND_MOUNT_BACKENDS.discard(name)
logger.info("Registered sandbox backend: %s (bind mounts: %s)", name, supports_bind_mounts)
def backend_supports_bind_mounts(name: str) -> bool:
return name in _BIND_MOUNT_BACKENDS
def supported_backends() -> list[str]:
+5 -5
View File
@@ -237,18 +237,18 @@ class StrixDockerSandboxClient(DockerSandboxClient):
_apply_log_limits(create_kwargs)
_apply_run_labels(create_kwargs)
# Strix injection: host bind mounts (e.g. large repos passed via --mount)
# that bypass the SDK's file-by-file LocalDir copy.
bind_mounts = getattr(self, "strix_bind_mounts", ())
# Strix injection: local source trees, sorted shallowest-first so a
# nested spec lands on top of the tree it covers.
bind_mounts = self.strix_bind_mounts or ()
if bind_mounts:
mounts = create_kwargs.setdefault("mounts", [])
for spec in bind_mounts:
for spec in sorted(bind_mounts, key=lambda s: str(s["target"]).count("/")):
mounts.append(
DockerSDKMount(
target=spec["target"],
source=spec["source"],
type="bind",
read_only=spec.get("read_only", True),
read_only=spec.get("read_only", False),
)
)
-120
View File
@@ -1,120 +0,0 @@
"""Symlink-safe staging for ``LocalDir`` manifest uploads.
The sandbox SDK's ``LocalDir`` walker refuses to copy symlinks at all — it
raises ``LocalDirReadError(reason="symlink_not_supported")`` on the first one
as a path-escape / TOCTOU safeguard. Real source trees (especially JS/TS
monorepos with workspace or shared-config links) routinely commit symlinks, so
handing such a tree straight to ``LocalDir`` aborts the upload before the agent
even starts.
:func:`stage_symlink_safe_dir` returns a path that is always safe to hand to
``LocalDir``:
* a tree with no symlinks is used as-is (no copy);
* otherwise the tree is copied into a temp directory with symlinks resolved:
- a link whose target stays inside the tree is *dereferenced* (its target
content is materialized in place), so the agent still sees the file;
- a link that escapes the tree, dangles, or forms a cycle is *dropped* and
never followed. Refusing to follow out-of-tree links preserves the walker's
path-escape safety and keeps host/out-of-tree content from leaking into the
(hostile) sandbox.
Regular files are hard-linked when possible (falling back to a copy across
devices), so the staged tree adds negligible disk for the non-symlink bulk.
"""
from __future__ import annotations
import logging
import os
import shutil
import tempfile
from pathlib import Path
logger = logging.getLogger(__name__)
_STAGING_PREFIX = "strix-localdir-"
def _is_within(target: Path, root: Path) -> bool:
"""Return whether ``target`` is ``root`` itself or nested under it."""
if target == root:
return True
try:
target.relative_to(root)
except ValueError:
return False
return True
def tree_has_symlink(root: Path) -> bool:
"""Return whether ``root`` contains any symlink (file or directory)."""
for dirpath, dirnames, filenames in os.walk(root, followlinks=False):
base = Path(dirpath)
for name in (*dirnames, *filenames):
if (base / name).is_symlink():
return True
return False
def _link_or_copy(src: Path, dst: Path) -> None:
"""Hard-link ``src`` to ``dst``, falling back to a content copy."""
try:
os.link(src, dst)
except OSError:
shutil.copy2(src, dst, follow_symlinks=True)
def _stage_dir(src: Path, dst: Path, root: Path, seen: frozenset[Path]) -> None:
dst.mkdir(parents=True, exist_ok=True)
for entry in os.scandir(src):
entry_path = Path(entry.path)
dest_path = dst / entry.name
if entry.is_symlink():
target = Path(os.path.realpath(entry_path))
if not _is_within(target, root):
logger.warning("staging: dropping out-of-tree symlink %s -> %s", entry_path, target)
continue
if not target.exists():
logger.warning("staging: dropping dangling symlink %s", entry_path)
continue
if target in seen:
logger.warning("staging: dropping cyclic symlink %s -> %s", entry_path, target)
continue
if target.is_dir():
_stage_dir(target, dest_path, root, seen | {target})
else:
_link_or_copy(target, dest_path)
elif entry.is_dir(follow_symlinks=False):
_stage_dir(entry_path, dest_path, root, seen)
elif entry.is_file(follow_symlinks=False):
_link_or_copy(entry_path, dest_path)
else:
# Sockets, FIFOs, devices — not part of a source tree; skip.
logger.debug("staging: skipping non-regular entry %s", entry_path)
def stage_symlink_safe_dir(src_root: Path) -> tuple[Path, Path | None]:
"""Return ``(upload_path, staged_temp)`` for uploading ``src_root``.
``upload_path`` is safe to hand to ``LocalDir``. When the tree contains no
symlinks it is ``src_root`` itself and ``staged_temp`` is ``None``.
Otherwise a symlink-safe copy is materialized in a temp directory and both
returned values point at it; the caller owns removing ``staged_temp`` once
the upload completes.
"""
root = src_root.resolve()
if not tree_has_symlink(root):
return root, None
staged = Path(tempfile.mkdtemp(prefix=_STAGING_PREFIX)).resolve()
try:
_stage_dir(root, staged, root, frozenset({root}))
except OSError:
shutil.rmtree(staged, ignore_errors=True)
raise
logger.info("staging: materialized symlink-safe copy of %s at %s", root, staged)
return staged, staged
+89 -47
View File
@@ -3,17 +3,21 @@
from __future__ import annotations
import logging
import shutil
import os
import sys
from pathlib import Path
from typing import Any
from typing import TYPE_CHECKING, Any
from agents.sandbox.entries import BaseEntry, LocalDir
from agents.sandbox.manifest import Environment, Manifest
from strix.config import load_settings
from strix.runtime.backends import get_backend
from strix.runtime.backends import backend_supports_bind_mounts, get_backend
from strix.runtime.caido_bootstrap import bootstrap_caido
from strix.runtime.local_dir_staging import stage_symlink_safe_dir
if TYPE_CHECKING:
from strix.runtime.status import StatusSink
logger = logging.getLogger(__name__)
@@ -28,43 +32,72 @@ _SESSION_CACHE: dict[str, dict[str, Any]] = {}
# Manifest root inside the container; entry keys hang off this path.
_WORKSPACE_ROOT = "/workspace"
_PROTECTED_METADATA_NAMES = (".git", ".agents", ".codex")
def build_session_entries(
local_sources: list[dict[str, Any]],
) -> tuple[dict[str | Path, BaseEntry], list[dict[str, Any]], list[Path]]:
"""Split local sources into copied manifest entries and host bind mounts.
Sources flagged ``mount`` are bind-mounted read-only at
``/workspace/<workspace_subdir>`` (not added to the manifest, so the SDK
does not stream them in file-by-file). Every other source becomes a
``LocalDir`` entry copied into the container as before. Trees containing
symlinks (which the SDK's ``LocalDir`` walker refuses outright) are first
staged into a symlink-safe temp copy; those temp dirs are returned so the
caller can remove them once the upload completes.
"""
entries: dict[str | Path, BaseEntry] = {}
def _host_identity_env() -> dict[str, str]:
if sys.platform != "linux":
return {}
return {"STRIX_HOST_UID": str(os.getuid()), "STRIX_HOST_GID": str(os.getgid())}
def build_bind_mounts(local_sources: list[dict[str, Any]]) -> list[dict[str, Any]]:
bind_mounts: list[dict[str, Any]] = []
staged_dirs: list[Path] = []
for src in local_sources:
ws_subdir = src.get("workspace_subdir") or ""
host_path = src.get("source_path") or ""
if not ws_subdir or not host_path:
continue
resolved = Path(host_path).expanduser().resolve()
if src.get("mount"):
bind_mounts.append(
{
"source": str(resolved),
"target": f"{_WORKSPACE_ROOT}/{ws_subdir}",
"read_only": True,
}
target = f"{_WORKSPACE_ROOT}/{ws_subdir}"
bind_mounts.append({"source": str(resolved), "target": target, "read_only": False})
if src.get("protect_metadata"):
bind_mounts.extend(_metadata_mounts(resolved, target))
return bind_mounts
def build_manifest_entries(local_sources: list[dict[str, Any]]) -> dict[str | Path, BaseEntry]:
entries: dict[str | Path, BaseEntry] = {}
for src in local_sources:
ws_subdir = src.get("workspace_subdir") or ""
host_path = src.get("source_path") or ""
if not ws_subdir or not host_path:
continue
entries[ws_subdir] = LocalDir(src=Path(host_path).expanduser().resolve())
return entries
def _metadata_mounts(tree: Path, target: str) -> list[dict[str, Any]]:
mounts: list[dict[str, Any]] = []
for name in _PROTECTED_METADATA_NAMES:
metadata = tree / name
if not metadata.is_dir() and not metadata.is_file():
continue
if not metadata.resolve().is_relative_to(tree):
continue
mounts.append({"source": str(metadata), "target": f"{target}/{name}", "read_only": True})
gitdir = _gitdir_from_pointer(metadata) if metadata.is_file() else None
if gitdir is not None and gitdir.exists() and gitdir.is_relative_to(tree):
relative = gitdir.relative_to(tree).as_posix()
mounts.append(
{"source": str(gitdir), "target": f"{target}/{relative}", "read_only": True}
)
else:
upload_path, staged = stage_symlink_safe_dir(resolved)
if staged is not None:
staged_dirs.append(staged)
entries[ws_subdir] = LocalDir(src=upload_path)
return entries, bind_mounts, staged_dirs
return mounts
def _gitdir_from_pointer(git_file: Path) -> Path | None:
try:
content = git_file.read_text(encoding="utf-8", errors="replace")
except OSError:
return None
for line in content.splitlines():
prefix, _, value = line.partition(":")
if prefix.strip() == "gitdir" and value.strip():
candidate = Path(value.strip()).expanduser()
if not candidate.is_absolute():
candidate = git_file.parent / candidate
return candidate.resolve()
return None
async def create_or_reuse(
@@ -72,19 +105,32 @@ async def create_or_reuse(
*,
image: str,
local_sources: list[dict[str, Any]],
status_sink: StatusSink | None = None,
) -> dict[str, Any]:
"""Return the existing session bundle for ``scan_id`` or create a new one.
Each ``local_sources`` entry exposes its host ``source_path`` at
``/workspace/<workspace_subdir>`` inside the container copied in, or
bind-mounted read-only when the entry is flagged ``mount``.
``/workspace/<workspace_subdir>`` inside the container.
"""
def report(phase: str) -> None:
if status_sink is not None:
status_sink(phase)
cached = _SESSION_CACHE.get(scan_id)
if cached is not None:
logger.info("Reusing existing sandbox session for scan %s", scan_id)
return cached
entries, bind_mounts, staged_dirs = build_session_entries(local_sources)
backend_name = load_settings().runtime.backend
backend = get_backend(backend_name)
if backend_supports_bind_mounts(backend_name):
bind_mounts = build_bind_mounts(local_sources)
entries: dict[str | Path, BaseEntry] = {}
else:
bind_mounts = []
entries = build_manifest_entries(local_sources)
# Caido runs as an in-container sidecar; HTTP(S) traffic from any
# process started via ``session.exec`` (the SDK's Shell tool, etc.)
@@ -98,6 +144,7 @@ async def create_or_reuse(
value={
"PYTHONUNBUFFERED": "1",
"HOST_GATEWAY": "host.docker.internal",
**_host_identity_env(),
"http_proxy": container_caido_url,
"https_proxy": container_caido_url,
"ALL_PROXY": container_caido_url,
@@ -106,26 +153,21 @@ async def create_or_reuse(
),
)
backend_name = load_settings().runtime.backend
backend = get_backend(backend_name)
logger.info(
"Creating sandbox session for scan %s (backend=%s, image=%s)",
scan_id,
backend_name,
image,
)
try:
client, session = await backend(
image=image,
manifest=manifest,
exposed_ports=(_CONTAINER_CAIDO_PORT,),
bind_mounts=bind_mounts,
)
finally:
for staged in staged_dirs:
shutil.rmtree(staged, ignore_errors=True)
report("Starting sandbox container")
client, session = await backend(
image=image,
manifest=manifest,
exposed_ports=(_CONTAINER_CAIDO_PORT,),
bind_mounts=bind_mounts,
)
report("Setting up the proxy")
caido_endpoint = await session.resolve_exposed_port(_CONTAINER_CAIDO_PORT)
scheme = "https" if caido_endpoint.tls else "http"
host_caido_url = f"{scheme}://{caido_endpoint.host}:{caido_endpoint.port}"
+8
View File
@@ -0,0 +1,8 @@
"""Startup phase reporting."""
from __future__ import annotations
from collections.abc import Callable
StatusSink = Callable[[str], None]
+15 -3
View File
@@ -13,6 +13,7 @@ from typing import Any, Literal, get_args
from agents import RunContextWrapper, function_tool
from strix.core.agents import Status, coordinator_from_context
from strix.core.execution import notify_parent_on_terminal
from strix.skills import validate_requested_skills
@@ -558,7 +559,7 @@ async def agent_finish(
)
parent_notified = False
if report_to_parent:
if report_to_parent and await coordinator.claim_parent_notice(me):
async with coordinator._lock:
agent_name = coordinator.names.get(me, me)
report = _render_completion_report(
@@ -582,6 +583,11 @@ async def agent_finish(
)
parent_notified = True
await coordinator.set_status(me, "completed")
if not parent_notified:
# Silence here would leave a parent waiting on a report that is never coming.
await notify_parent_on_terminal(coordinator, me, "completed")
logger.info(
"agent_finish: %s success=%s findings=%d parent_notified=%s",
me,
@@ -589,7 +595,6 @@ async def agent_finish(
len(findings or []),
parent_notified,
)
await coordinator.set_status(me, "completed")
return json.dumps(
{
@@ -680,9 +685,16 @@ async def stop_agent(
)
if cascade:
await coordinator.cancel_descendants_graceful(target_agent_id)
stopped = await coordinator.cancel_descendants_graceful(target_agent_id)
else:
await coordinator.request_stop(target_agent_id)
stopped = [target_agent_id]
# The stopper knows what it just did; anyone else waiting on those agents does not.
async with coordinator._lock:
orphaned = [aid for aid in stopped if coordinator.parent_of.get(aid) not in (None, me)]
for aid in orphaned:
await notify_parent_on_terminal(coordinator, aid, "stopped")
logger.info(
"stop_agent: target=%s cascade=%s reason=%r",
+2 -7
View File
@@ -30,9 +30,7 @@ def test_parse_arguments_accepts_target_list_file(
) -> None:
target_list = tmp_path / "targets.txt"
target_list.write_text(
"https://test1.com/\n"
"\n"
"http://test2.com:5789/\n",
"https://test1.com/\n\nhttp://test2.com:5789/\n",
encoding="utf-8",
)
_stub_settings(monkeypatch)
@@ -84,7 +82,4 @@ def test_parse_arguments_rejects_resume_with_target_list(
with pytest.raises(SystemExit):
cli_main.parse_arguments()
assert (
"Cannot combine --resume with --target/--target-list/--mount"
in capsys.readouterr().err
)
assert "Cannot combine --resume with --target/--target-list" in capsys.readouterr().err
-1
View File
@@ -33,7 +33,6 @@ _LLM_ENV_KEYS = [
# RuntimeSettings
"STRIX_IMAGE",
"STRIX_RUNTIME_BACKEND",
"STRIX_MAX_LOCAL_COPY_MB",
# TelemetrySettings
"STRIX_TELEMETRY",
]
+115 -8
View File
@@ -18,10 +18,11 @@ from openai.types.responses import ResponseOutputMessage, ResponseOutputRefusal
from strix.core import execution
from strix.core.agents import AgentCoordinator
from strix.core.execution import (
_notify_parent_on_terminal,
_notify_root_on_budget_reserve,
notify_parent_on_terminal,
)
from strix.core.sessions import seed_initial_input
from strix.tools.agents_graph.tools import agent_finish, stop_agent
from strix.tools.finish.tool import finish_scan
@@ -67,6 +68,43 @@ async def _call_finish_scan(
return parsed
async def _call_agent_finish(
coordinator: AgentCoordinator,
agent_id: str,
parent_id: str | None,
*,
report_to_parent: bool,
) -> dict[str, Any]:
ctx = ToolContext(
context={"coordinator": coordinator, "agent_id": agent_id, "parent_id": parent_id},
tool_name="agent_finish",
tool_call_id="call-1",
tool_arguments="{}",
)
result: str = await agent_finish.on_invoke_tool(
ctx,
json.dumps({"result_summary": "done", "report_to_parent": report_to_parent}),
)
parsed: dict[str, Any] = json.loads(result)
return parsed
async def _call_stop_agent(
coordinator: AgentCoordinator, agent_id: str, target_agent_id: str
) -> dict[str, Any]:
ctx = ToolContext(
context={"coordinator": coordinator, "agent_id": agent_id},
tool_name="stop_agent",
tool_call_id="call-1",
tool_arguments="{}",
)
result: str = await stop_agent.on_invoke_tool(
ctx, json.dumps({"target_agent_id": target_agent_id})
)
parsed: dict[str, Any] = json.loads(result)
return parsed
@pytest.mark.asyncio
async def test_reserve_stop_notifies_root_once(monkeypatch: pytest.MonkeyPatch) -> None:
coordinator = AgentCoordinator()
@@ -466,11 +504,11 @@ async def test_snapshot_round_trip_preserves_budget_pause() -> None:
@pytest.mark.asyncio
@pytest.mark.parametrize("status", ["stopped", "failed", "crashed"])
@pytest.mark.parametrize("status", ["completed", "stopped", "failed", "crashed"])
async def test_terminal_child_wakes_parked_parent(tmp_path: Any, status: str) -> None:
# Regression for #870: a child reaching a terminal state (e.g. MaxTurnsExceeded
# -> "stopped") must wake the parent parked in wait_for_message, so the root can
# finalize the scan instead of hanging for a completion report that never arrives.
# Regression for #870 and #947: a child reaching any terminal state - including a
# plain "completed" - must wake the parent parked in wait_for_agents, so the root
# can finalize the scan instead of hanging for a report that never arrives.
coordinator = AgentCoordinator()
await coordinator.register("root", "strix", parent_id=None)
await coordinator.register("child", "SQL Injection", parent_id="root")
@@ -482,13 +520,82 @@ async def test_terminal_child_wakes_parked_parent(tmp_path: Any, status: str) ->
assert not root_waiter.done()
await coordinator.set_status("child", status, error="Max turns (500) exceeded")
await _notify_parent_on_terminal(coordinator, "child", status)
await notify_parent_on_terminal(coordinator, "child", status)
await asyncio.wait_for(root_waiter, timeout=1.0)
assert coordinator.pending_counts.get("root", 0) > 0
session.close()
@pytest.mark.asyncio
async def test_agent_finish_without_report_still_wakes_parent(tmp_path: Any) -> None:
# Regression for #947: a child that completes with report_to_parent=False owes its
# parent a terminal notice, otherwise the parent waits out its full timeout.
coordinator = AgentCoordinator()
await coordinator.register("root", "strix", parent_id=None)
await coordinator.register("child", "recon", parent_id="root")
session = SQLiteSession("root", tmp_path / "agents.db")
await coordinator.attach_runtime("root", session=session)
root_waiter = asyncio.create_task(coordinator.wait_for_message("root"))
await asyncio.sleep(0)
await _call_agent_finish(coordinator, "child", "root", report_to_parent=False)
await asyncio.wait_for(root_waiter, timeout=1.0)
assert coordinator.statuses["child"] == "completed"
assert coordinator.pending_counts.get("root", 0) == 1
session.close()
@pytest.mark.asyncio
async def test_agent_finish_report_suppresses_the_terminal_notice(tmp_path: Any) -> None:
coordinator = AgentCoordinator()
await coordinator.register("root", "strix", parent_id=None)
await coordinator.register("child", "recon", parent_id="root")
session = SQLiteSession("root", tmp_path / "agents.db")
await coordinator.attach_runtime("root", session=session)
await _call_agent_finish(coordinator, "child", "root", report_to_parent=True)
# The exit backstop must not duplicate the report the child already delivered.
await execution._notify_parent_on_exit(coordinator, "child")
assert coordinator.pending_counts.get("root", 0) == 1
session.close()
@pytest.mark.asyncio
async def test_stop_agent_notifies_a_parent_that_is_not_the_stopper(tmp_path: Any) -> None:
coordinator = AgentCoordinator()
await coordinator.register("root", "strix", parent_id=None)
await coordinator.register("child", "recon", parent_id="root")
await coordinator.register("grandchild", "sqli", parent_id="child")
session = SQLiteSession("child", tmp_path / "agents.db")
await coordinator.attach_runtime("child", session=session)
await _call_stop_agent(coordinator, "root", "grandchild")
assert coordinator.statuses["grandchild"] == "stopped"
assert coordinator.pending_counts.get("child", 0) == 1
# The stopper already knows; only the waiting parent needs telling.
assert coordinator.pending_counts.get("root", 0) == 0
session.close()
@pytest.mark.asyncio
async def test_stop_agent_does_not_notify_the_stopping_parent(tmp_path: Any) -> None:
coordinator = AgentCoordinator()
await coordinator.register("root", "strix", parent_id=None)
await coordinator.register("child", "recon", parent_id="root")
session = SQLiteSession("root", tmp_path / "agents.db")
await coordinator.attach_runtime("root", session=session)
await _call_stop_agent(coordinator, "root", "child")
assert coordinator.pending_counts.get("root", 0) == 0
session.close()
@pytest.mark.asyncio
async def test_notify_parent_on_terminal_ignores_non_terminal_status(tmp_path: Any) -> None:
coordinator = AgentCoordinator()
@@ -497,7 +604,7 @@ async def test_notify_parent_on_terminal_ignores_non_terminal_status(tmp_path: A
session = SQLiteSession("root", tmp_path / "agents.db")
await coordinator.attach_runtime("root", session=session)
await _notify_parent_on_terminal(coordinator, "child", "waiting")
await notify_parent_on_terminal(coordinator, "child", "waiting")
assert coordinator.pending_counts.get("root", 0) == 0
session.close()
@@ -523,7 +630,7 @@ async def test_terminal_notice_does_not_cancel_parent_stream(tmp_path: Any) -> N
await coordinator.attach_runtime("root", session=session, interrupt_on_message=True)
await coordinator.attach_stream("root", stream)
await _notify_parent_on_terminal(coordinator, "child", "crashed")
await notify_parent_on_terminal(coordinator, "child", "crashed")
assert stream.cancelled is False
assert coordinator.pending_counts.get("root", 0) > 0
-143
View File
@@ -1,143 +0,0 @@
"""Tests for symlink-safe LocalDir staging."""
from __future__ import annotations
from typing import TYPE_CHECKING
from strix.runtime.local_dir_staging import stage_symlink_safe_dir, tree_has_symlink
if TYPE_CHECKING:
from pathlib import Path
def _make_repo(tmp_path: Path) -> Path:
repo = tmp_path / "repo"
(repo / "pkg").mkdir(parents=True)
(repo / "pkg" / "mod.py").write_text("x = 1\n")
(repo / "README.md").write_text("readme\n")
return repo
def test_tree_without_symlinks_used_as_is(tmp_path: Path) -> None:
repo = _make_repo(tmp_path)
upload_path, staged = stage_symlink_safe_dir(repo)
assert staged is None
assert upload_path == repo.resolve()
assert not tree_has_symlink(repo)
def test_in_tree_file_symlink_is_dereferenced(tmp_path: Path) -> None:
repo = _make_repo(tmp_path)
(repo / "link.py").symlink_to(repo / "pkg" / "mod.py")
upload_path, staged = stage_symlink_safe_dir(repo)
assert staged is not None
assert upload_path == staged
assert not (staged / "link.py").is_symlink()
assert (staged / "link.py").read_text() == "x = 1\n"
assert (staged / "pkg" / "mod.py").read_text() == "x = 1\n"
assert not tree_has_symlink(staged)
def test_in_tree_relative_dir_symlink_is_dereferenced(tmp_path: Path) -> None:
repo = _make_repo(tmp_path)
(repo / "pkg_alias").symlink_to("pkg")
_upload, staged = stage_symlink_safe_dir(repo)
assert staged is not None
assert (staged / "pkg_alias" / "mod.py").read_text() == "x = 1\n"
assert not tree_has_symlink(staged)
def test_out_of_tree_symlink_is_dropped(tmp_path: Path) -> None:
repo = _make_repo(tmp_path)
outside = tmp_path / "outside.txt"
outside.write_text("secret\n")
(repo / "escape.txt").symlink_to(outside)
(repo / "abs_escape").symlink_to("/etc")
_upload, staged = stage_symlink_safe_dir(repo)
assert staged is not None
assert not (staged / "escape.txt").exists()
assert not (staged / "abs_escape").exists()
assert (staged / "README.md").exists()
def test_dangling_symlink_is_dropped(tmp_path: Path) -> None:
repo = _make_repo(tmp_path)
(repo / "dangling").symlink_to(repo / "does-not-exist")
_upload, staged = stage_symlink_safe_dir(repo)
assert staged is not None
assert not (staged / "dangling").exists()
assert not (staged / "dangling").is_symlink()
def test_cyclic_symlink_terminates(tmp_path: Path) -> None:
repo = _make_repo(tmp_path)
(repo / "self").symlink_to(repo)
(repo / "pkg" / "up").symlink_to("..")
_upload, staged = stage_symlink_safe_dir(repo)
assert staged is not None
assert (staged / "README.md").exists()
assert not tree_has_symlink(staged)
def test_nested_symlinks_inside_linked_dir(tmp_path: Path) -> None:
repo = _make_repo(tmp_path)
shared = repo / "shared"
shared.mkdir()
(shared / "conf.json").write_text("{}\n")
(shared / "escape").symlink_to("/etc/passwd")
(repo / "pkg" / "shared_link").symlink_to(shared)
_upload, staged = stage_symlink_safe_dir(repo)
assert staged is not None
assert (staged / "pkg" / "shared_link" / "conf.json").read_text() == "{}\n"
assert not (staged / "pkg" / "shared_link" / "escape").exists()
assert not (staged / "shared" / "escape").exists()
def test_staged_path_has_no_symlink_ancestor(tmp_path: Path, monkeypatch) -> None: # noqa: ANN001
"""The staging directory itself must never sit behind a symlink.
``tempfile.mkdtemp()`` honors ``$TMPDIR``, and on macOS the default
``$TMPDIR`` resolves through ``/var``, which is itself a symlink to
``/private/var``. ``LocalDir`` rejects any symlink component in its
source path, so returning the raw ``mkdtemp()`` result breaks every
local-dir upload on macOS whenever the source tree contains a symlink.
This reproduces that shape without depending on the host OS layout.
"""
repo = _make_repo(tmp_path)
(repo / "link.py").symlink_to(repo / "pkg" / "mod.py")
real_tmp_root = tmp_path / "real_tmp"
real_tmp_root.mkdir()
symlinked_tmp_root = tmp_path / "tmp_symlink"
symlinked_tmp_root.symlink_to(real_tmp_root)
def fake_mkdtemp(prefix: str = "") -> str:
real_dir = real_tmp_root / f"{prefix}fake"
real_dir.mkdir()
return str(symlinked_tmp_root / real_dir.name)
monkeypatch.setattr(
"strix.runtime.local_dir_staging.tempfile.mkdtemp", fake_mkdtemp
)
upload_path, staged = stage_symlink_safe_dir(repo)
assert staged is not None
assert upload_path == staged
for path in (staged, *staged.parents):
assert not path.is_symlink(), f"staged path has a symlink ancestor: {path}"
+105 -157
View File
@@ -1,171 +1,138 @@
"""Tests for local-source sizing and ``--mount`` target helpers in interface.utils."""
"""Tests for local-source collection and mount policy in interface.utils."""
from __future__ import annotations
import logging
import os
import sys
from typing import TYPE_CHECKING, Any
from pathlib import Path
from typing import Any
import pytest
if TYPE_CHECKING:
from pathlib import Path
from strix.interface.utils import (
build_mount_targets_info,
check_mountable_dir,
collect_local_sources,
dedupe_local_targets,
directory_size_bytes,
find_oversized_local_targets,
infer_target_type,
read_target_list_file,
)
def _write_file(path: Path, size: int) -> None:
path.write_bytes(b"x" * size)
def _local_target(target_path: str) -> dict[str, Any]:
return {
"type": "local_code",
"details": {"target_path": target_path, "workspace_subdir": "repo"},
"original": target_path,
}
def _local_target(target_path: str, *, mount: bool = False) -> dict[str, Any]:
details: dict[str, Any] = {"target_path": target_path, "workspace_subdir": "repo"}
if mount:
details["mount"] = True
return {"type": "local_code", "details": details, "original": target_path}
def test_collect_local_sources_protects_the_users_own_git() -> None:
sources = collect_local_sources([_local_target("/code")])
assert sources == [
{"source_path": "/code", "workspace_subdir": "repo", "protect_metadata": True}
]
def test_directory_size_empty_dir_is_zero(tmp_path: Path) -> None:
assert directory_size_bytes(tmp_path) == 0
def test_directory_size_sums_flat_and_nested_files(tmp_path: Path) -> None:
_write_file(tmp_path / "a.txt", 100)
nested = tmp_path / "sub" / "deep"
nested.mkdir(parents=True)
_write_file(nested / "b.txt", 250)
assert directory_size_bytes(tmp_path) == 350
def test_directory_size_skips_symlinks(tmp_path: Path) -> None:
_write_file(tmp_path / "real.txt", 100)
(tmp_path / "link.txt").symlink_to(tmp_path / "real.txt")
# The symlink target is counted once via the real file, not doubled.
assert directory_size_bytes(tmp_path) == 100
@pytest.mark.skipif(sys.platform == "win32", reason="relies on POSIX permissions")
def test_directory_size_logs_and_skips_unreadable_subdir(
tmp_path: Path, caplog: pytest.LogCaptureFixture
) -> None:
if hasattr(os, "geteuid") and os.geteuid() == 0:
pytest.skip("root bypasses directory permissions")
_write_file(tmp_path / "top.txt", 100)
locked = tmp_path / "locked"
locked.mkdir()
_write_file(locked / "secret.bin", 9999)
locked.chmod(0o000)
try:
with caplog.at_level(logging.WARNING):
size = directory_size_bytes(tmp_path)
finally:
locked.chmod(0o755)
# The unreadable subtree is excluded (not silently treated as readable) and
# the omission is logged rather than vanishing without a trace.
assert size == 100
assert any("Could not read" in record.message for record in caplog.records)
def test_find_oversized_returns_nothing_under_limit(tmp_path: Path) -> None:
_write_file(tmp_path / "a.txt", 100)
targets = [_local_target(str(tmp_path))]
assert find_oversized_local_targets(targets, max_bytes=1000) == []
def test_find_oversized_returns_target_over_limit(tmp_path: Path) -> None:
_write_file(tmp_path / "big.bin", 500)
targets = [_local_target(str(tmp_path))]
result = find_oversized_local_targets(targets, max_bytes=100)
assert result == [(str(tmp_path), 500)]
def test_find_oversized_ignores_mounted_targets(tmp_path: Path) -> None:
_write_file(tmp_path / "big.bin", 500)
targets = [_local_target(str(tmp_path), mount=True)]
assert find_oversized_local_targets(targets, max_bytes=100) == []
def test_find_oversized_ignores_non_local_targets() -> None:
targets = [{"type": "web_application", "details": {"target_url": "https://x"}}]
assert find_oversized_local_targets(targets, max_bytes=1) == []
@pytest.mark.parametrize("disabled", [0, -1])
def test_find_oversized_disabled_for_non_positive_limit(tmp_path: Path, disabled: int) -> None:
_write_file(tmp_path / "big.bin", 500)
targets = [_local_target(str(tmp_path))]
assert find_oversized_local_targets(targets, max_bytes=disabled) == []
def test_collect_local_sources_propagates_mount_flag() -> None:
copied = _local_target("/copied")
copied["details"]["workspace_subdir"] = "copied"
mounted = _local_target("/mounted", mount=True)
mounted["details"]["workspace_subdir"] = "mounted"
sources = collect_local_sources([copied, mounted])
by_path = {s["source_path"]: s for s in sources}
assert by_path["/copied"]["mount"] is False
assert by_path["/mounted"]["mount"] is True
def test_collect_local_sources_repository_is_never_mounted() -> None:
def test_collect_local_sources_leaves_a_clone_writable() -> None:
repo = {
"type": "repository",
"details": {"cloned_repo_path": "/clone", "workspace_subdir": "clone"},
}
sources = collect_local_sources([repo])
assert sources == [{"source_path": "/clone", "workspace_subdir": "clone", "mount": False}]
assert sources == [
{"source_path": "/clone", "workspace_subdir": "clone", "protect_metadata": False}
]
def test_build_mount_targets_info_for_valid_dir(tmp_path: Path) -> None:
result = build_mount_targets_info([str(tmp_path)])
assert len(result) == 1
entry = result[0]
assert entry["type"] == "local_code"
assert entry["details"]["mount"] is True
assert entry["details"]["target_path"] == str(tmp_path.resolve())
def test_check_mountable_dir_accepts_a_project_dir(tmp_path: Path) -> None:
check_mountable_dir(tmp_path)
def test_build_mount_targets_info_rejects_missing_path(tmp_path: Path) -> None:
missing = tmp_path / "does-not-exist"
def test_check_mountable_dir_rejects_missing_path(tmp_path: Path) -> None:
with pytest.raises(ValueError, match="not an existing directory"):
build_mount_targets_info([str(missing)])
check_mountable_dir(tmp_path / "nope")
def test_build_mount_targets_info_rejects_file(tmp_path: Path) -> None:
file_path = tmp_path / "a-file.txt"
_write_file(file_path, 10)
with pytest.raises(ValueError, match="not an existing directory"):
build_mount_targets_info([str(file_path)])
def test_check_mountable_dir_rejects_filesystem_root() -> None:
with pytest.raises(ValueError, match="Refusing to mount"):
check_mountable_dir(Path("/"))
@pytest.mark.parametrize("empty", ["", " "])
def test_build_mount_targets_info_rejects_empty_path(empty: str) -> None:
# An empty path would otherwise resolve to the current working directory
# and silently bind-mount it into the sandbox.
with pytest.raises(ValueError, match="must not be empty"):
build_mount_targets_info([empty])
def test_check_mountable_dir_rejects_home(tmp_path: Path, monkeypatch: pytest.MonkeyPatch) -> None:
home = tmp_path / "home"
home.mkdir()
monkeypatch.setenv("HOME", str(home))
monkeypatch.setattr(Path, "home", classmethod(lambda _cls: home))
with pytest.raises(ValueError, match="Refusing to mount"):
check_mountable_dir(home)
def test_check_mountable_dir_rejects_system_root() -> None:
etc = Path("/etc")
if not etc.is_dir():
pytest.skip("no /etc on this platform")
with pytest.raises(ValueError, match="Refusing to mount"):
check_mountable_dir(etc)
def test_check_mountable_dir_rejects_the_shared_home_root() -> None:
home_root = Path("/home")
if not home_root.is_dir():
pytest.skip("no /home on this platform")
with pytest.raises(ValueError, match="Refusing to mount"):
check_mountable_dir(home_root)
def test_check_mountable_dir_matches_forbidden_names_case_insensitively(tmp_path: Path) -> None:
ssh_dir = tmp_path / ".SSH"
ssh_dir.mkdir()
with pytest.raises(ValueError, match="holds credentials"):
check_mountable_dir(ssh_dir)
def test_check_mountable_dir_rejects_credential_dirs(tmp_path: Path) -> None:
ssh_dir = tmp_path / ".ssh"
ssh_dir.mkdir()
with pytest.raises(ValueError, match="holds credentials"):
check_mountable_dir(ssh_dir)
def test_check_mountable_dir_rejects_credential_subdirs(tmp_path: Path) -> None:
keys = tmp_path / ".ssh" / "keys"
keys.mkdir(parents=True)
with pytest.raises(ValueError, match="holds credentials"):
check_mountable_dir(keys)
def test_check_mountable_dir_rejects_system_subdirs() -> None:
system_subdir = next((p for p in (Path("/etc/ssl"), Path("/usr/bin")) if p.is_dir()), None)
if system_subdir is None:
pytest.skip("no system subdirectory on this platform")
with pytest.raises(ValueError, match="Refusing to mount"):
check_mountable_dir(system_subdir)
def test_check_mountable_dir_accepts_a_project_under_the_home_root(
tmp_path: Path, monkeypatch: pytest.MonkeyPatch
) -> None:
project = tmp_path / "home" / "dev" / "project"
project.mkdir(parents=True)
monkeypatch.setattr(Path, "home", classmethod(lambda _cls: tmp_path / "home" / "dev"))
check_mountable_dir(project)
def test_infer_target_type_applies_the_mount_policy() -> None:
with pytest.raises(ValueError, match="Refusing to mount"):
infer_target_type("/etc")
def test_read_target_list_file_strips_blank_lines(tmp_path: Path) -> None:
target_list = tmp_path / "targets.txt"
target_list.write_text(
"\n"
" https://test1.com/ \n"
"\n"
"http://test2.com:5789/\n"
" \n",
"\n https://test1.com/ \n\nhttp://test2.com:5789/\n \n",
encoding="utf-8",
)
@@ -178,10 +145,7 @@ def test_read_target_list_file_strips_blank_lines(tmp_path: Path) -> None:
def test_read_target_list_file_ignores_comment_lines(tmp_path: Path) -> None:
target_list = tmp_path / "targets.txt"
target_list.write_text(
"# production targets\n"
"https://test1.com/\n"
" # staging targets\n"
"http://test2.com:5789/\n",
"# production targets\nhttps://test1.com/\n # staging targets\nhttp://test2.com:5789/\n",
encoding="utf-8",
)
@@ -222,28 +186,12 @@ def test_dedupe_keeps_distinct_targets_in_order() -> None:
targets = [
_local_target("/a"),
{"type": "web_application", "details": {"target_url": "https://x"}},
_local_target("/b", mount=True),
_local_target("/b"),
]
assert dedupe_local_targets(targets) == targets
def test_dedupe_mount_supersedes_copied_same_path() -> None:
copied = _local_target("/repo")
mounted = _local_target("/repo", mount=True)
# Copied first, then mounted: the single surviving entry is the mount.
result = dedupe_local_targets([copied, mounted])
assert len(result) == 1
assert result[0]["details"]["mount"] is True
# Order-independent: mounted first, copied second also yields the mount.
result_rev = dedupe_local_targets([mounted, copied])
assert len(result_rev) == 1
assert result_rev[0]["details"]["mount"] is True
def test_dedupe_collapses_duplicate_mounts() -> None:
result = dedupe_local_targets(
[_local_target("/repo", mount=True), _local_target("/repo", mount=True)]
)
assert len(result) == 1
def test_dedupe_collapses_the_same_path() -> None:
assert dedupe_local_targets([_local_target("/repo"), _local_target("/repo")]) == [
_local_target("/repo")
]
+147 -54
View File
@@ -1,4 +1,4 @@
"""Tests for build_session_entries: splitting copied vs bind-mounted sources."""
"""Tests for how local sources reach the sandbox: bind mounts or manifest upload."""
from __future__ import annotations
@@ -6,82 +6,175 @@ from typing import TYPE_CHECKING, Any
from agents.sandbox.entries import LocalDir
from strix.runtime.session_manager import build_session_entries
from strix.runtime.backends import (
_BACKENDS,
_BIND_MOUNT_BACKENDS,
backend_supports_bind_mounts,
register_backend,
)
from strix.runtime.session_manager import build_bind_mounts, build_manifest_entries
if TYPE_CHECKING:
from pathlib import Path
def _source(subdir: str, path: str, *, mount: bool = False) -> dict[str, Any]:
return {"source_path": path, "workspace_subdir": subdir, "mount": mount}
def _source(subdir: str, path: str, *, protect_metadata: bool = False) -> dict[str, Any]:
return {"source_path": path, "workspace_subdir": subdir, "protect_metadata": protect_metadata}
def test_copied_source_becomes_localdir_entry(tmp_path: Path) -> None:
entries, bind_mounts, staged_dirs = build_session_entries([_source("repo", str(tmp_path))])
assert bind_mounts == []
assert staged_dirs == []
assert isinstance(entries["repo"], LocalDir)
assert entries["repo"].src == tmp_path.resolve()
def test_mounted_source_becomes_bind_mount(tmp_path: Path) -> None:
entries, bind_mounts, _staged = build_session_entries(
[_source("repo", str(tmp_path), mount=True)]
)
assert entries == {}
assert bind_mounts == [
def test_source_becomes_writable_bind_mount(tmp_path: Path) -> None:
assert build_bind_mounts([_source("repo", str(tmp_path))]) == [
{
"source": str(tmp_path.resolve()),
"target": "/workspace/repo",
"read_only": True,
"read_only": False,
}
]
def test_mixed_sources_split_correctly(tmp_path: Path) -> None:
copied = tmp_path / "copied"
mounted = tmp_path / "mounted"
copied.mkdir()
mounted.mkdir()
def test_git_dir_is_remounted_read_only_when_protected(tmp_path: Path) -> None:
(tmp_path / ".git").mkdir()
entries, bind_mounts, _staged = build_session_entries(
[
_source("copied", str(copied)),
_source("mounted", str(mounted), mount=True),
]
)
mounts = build_bind_mounts([_source("repo", str(tmp_path), protect_metadata=True)])
assert list(entries) == ["copied"]
assert isinstance(entries["copied"], LocalDir)
assert [m["target"] for m in bind_mounts] == ["/workspace/mounted"]
assert mounts == [
{"source": str(tmp_path.resolve()), "target": "/workspace/repo", "read_only": False},
{
"source": str((tmp_path / ".git").resolve()),
"target": "/workspace/repo/.git",
"read_only": True,
},
]
def test_agent_instruction_dirs_are_protected_too(tmp_path: Path) -> None:
(tmp_path / ".agents").mkdir()
(tmp_path / ".codex").mkdir()
mounts = build_bind_mounts([_source("repo", str(tmp_path), protect_metadata=True)])
assert [(m["target"], m["read_only"]) for m in mounts] == [
("/workspace/repo", False),
("/workspace/repo/.agents", True),
("/workspace/repo/.codex", True),
]
def test_worktree_git_pointer_file_is_protected(tmp_path: Path) -> None:
gitdir = tmp_path / "nested" / "gitdir"
gitdir.mkdir(parents=True)
(tmp_path / ".git").write_text(f"gitdir: {gitdir}\n", encoding="utf-8")
mounts = build_bind_mounts([_source("repo", str(tmp_path), protect_metadata=True)])
assert [(m["target"], m["read_only"]) for m in mounts] == [
("/workspace/repo", False),
("/workspace/repo/.git", True),
("/workspace/repo/nested/gitdir", True),
]
def test_git_pointer_to_a_missing_gitdir_is_not_mounted(tmp_path: Path) -> None:
(tmp_path / ".git").write_text(f"gitdir: {tmp_path / 'gone'}\n", encoding="utf-8")
mounts = build_bind_mounts([_source("repo", str(tmp_path), protect_metadata=True)])
assert [m["target"] for m in mounts] == ["/workspace/repo", "/workspace/repo/.git"]
def test_git_pointer_outside_the_tree_needs_no_nested_mount(tmp_path: Path) -> None:
tree = tmp_path / "worktree"
tree.mkdir()
(tree / ".git").write_text(f"gitdir: {tmp_path / 'main' / '.git'}\n", encoding="utf-8")
mounts = build_bind_mounts([_source("repo", str(tree), protect_metadata=True)])
assert [m["target"] for m in mounts] == ["/workspace/repo", "/workspace/repo/.git"]
def test_metadata_symlinked_outside_the_tree_is_not_mounted(tmp_path: Path) -> None:
outside = tmp_path / "elsewhere"
outside.mkdir()
tree = tmp_path / "repo"
tree.mkdir()
(tree / ".git").symlink_to(outside, target_is_directory=True)
mounts = build_bind_mounts([_source("repo", str(tree), protect_metadata=True)])
assert [m["target"] for m in mounts] == ["/workspace/repo"]
def test_no_git_guard_without_a_git_dir(tmp_path: Path) -> None:
mounts = build_bind_mounts([_source("repo", str(tmp_path), protect_metadata=True)])
assert [m["target"] for m in mounts] == ["/workspace/repo"]
def test_clone_keeps_its_git_writable(tmp_path: Path) -> None:
(tmp_path / ".git").mkdir()
mounts = build_bind_mounts([_source("clone", str(tmp_path), protect_metadata=False)])
assert [m["target"] for m in mounts] == ["/workspace/clone"]
def test_multiple_sources_each_get_a_mount(tmp_path: Path) -> None:
first = tmp_path / "first"
second = tmp_path / "second"
first.mkdir()
second.mkdir()
mounts = build_bind_mounts([_source("first", str(first)), _source("second", str(second))])
assert [m["target"] for m in mounts] == ["/workspace/first", "/workspace/second"]
assert all(m["read_only"] is False for m in mounts)
def test_incomplete_sources_are_skipped() -> None:
entries, bind_mounts, staged_dirs = build_session_entries(
[
{"source_path": "", "workspace_subdir": "x"},
{"source_path": "/p", "workspace_subdir": ""},
]
assert (
build_bind_mounts(
[
{"source_path": "", "workspace_subdir": "x"},
{"source_path": "/p", "workspace_subdir": ""},
]
)
== []
)
assert entries == {}
assert bind_mounts == []
assert staged_dirs == []
def test_symlink_tree_is_staged(tmp_path: Path) -> None:
repo = tmp_path / "repo"
repo.mkdir()
(repo / "real.txt").write_text("content")
(repo / "link.txt").symlink_to(repo / "real.txt")
def test_manifest_entries_upload_sources_for_backends_without_bind_mounts(
tmp_path: Path,
) -> None:
entries = build_manifest_entries([_source("repo", str(tmp_path), protect_metadata=True)])
entries, _mounts, staged_dirs = build_session_entries([_source("repo", str(repo))])
assert len(staged_dirs) == 1
assert set(entries) == {"repo"}
entry = entries["repo"]
assert isinstance(entry, LocalDir)
assert entry.src == staged_dirs[0]
assert not (staged_dirs[0] / "link.txt").is_symlink()
assert (staged_dirs[0] / "link.txt").read_text() == "content"
assert entry.src == tmp_path.resolve()
def test_manifest_entries_skip_incomplete_sources() -> None:
assert (
build_manifest_entries(
[
{"source_path": "", "workspace_subdir": "x"},
{"source_path": "/p", "workspace_subdir": ""},
]
)
== {}
)
def test_only_bind_mount_capable_backends_are_registered_as_such() -> None:
assert backend_supports_bind_mounts("docker")
assert not backend_supports_bind_mounts("e2b")
async def _remote_backend(**_kwargs: Any) -> tuple[Any, Any]:
return object(), object()
try:
register_backend("e2b", _remote_backend)
assert not backend_supports_bind_mounts("e2b")
register_backend("e2b", _remote_backend, supports_bind_mounts=True)
assert backend_supports_bind_mounts("e2b")
finally:
_BACKENDS.pop("e2b", None)
_BIND_MOUNT_BACKENDS.discard("e2b")