SandboxResult

Run commands and agent-written programs inside an operating-system boundary. See the guide.

class zett_agent.sandbox.base.SandboxResult(exit_code, stdout, stderr, timed_out=False, stdout_truncated=False, stderr_truncated=False, duration=0.0, sandbox='', isolated=False, warnings=(), resources=ResourceReport(memory='none', cpu_seconds='none', cpu_percent='none', processes='none', file_size_bytes='none', open_files='none', core_dumps='none', peak_memory_bytes=None, limits_hit=(), warnings=()))

What one sandboxed run produced.

参数:

• exit_code (int | None)

• stdout (str)

• stderr (str)

• timed_out (bool)

• stdout_truncated (bool)

• stderr_truncated (bool)

• duration (float)

• sandbox (str)

• isolated (bool)

• warnings (tuple[str, ...])

• resources (ResourceReport)

exit_code

The process's exit status. A negative number is the signal that ended it, which is what a timeout produces. Bubblewrap runs the command under its own init, which reports a signal as 128 plus its number (137 for SIGKILL), as a shell does.

Type:

int | None

stdout

Standard output, decoded as UTF-8 with undecodable bytes replaced.

Type:

str

stderr

Standard error, decoded the same way.

Type:

str

timed_out

Whether the run exceeded its time limit and was killed.

Type:

bool

stdout_truncated

Whether standard output exceeded the byte limit.

Type:

bool

stderr_truncated

Whether standard error exceeded the byte limit.

Type:

bool

duration

Seconds the run took.

Type:

float

sandbox

Name of the backend that ran it.

Type:

str

isolated

Whether that backend enforces a boundary at all.

Type:

bool

warnings

Parts of the policy the backend could not enforce. Resource warnings are on resources instead.

Type:

tuple[str, ...]

resources

How each limit was enforced, what the run measured, and the limits it could not enforce here.

Type:

zett_agent.sandbox.resources.ResourceReport

property succeeded: bool

Whether the process exited with status 0 within its time limit.