II.
ScopeBoundary overview
Reference · livescope-boundary:terminal-bench.scope
scope-boundary:terminal-bench.scope overview
Inspect the raw attributes, linked wiki pages, and inbound or outbound graph edges for scope-boundary:terminal-bench.scope.
Attributes
subjectId
inScope
Terminal-task benchmark — agent operates a real shell to complete sysadmin / dev-tooling tasks; scored by per-task success and step-budget.
outOfScope
GUI-based agents, browser-only tasks (use WebArena), and benchmarks without a real-shell harness.
outOfScopeReasonIds
Outgoing edges
bounds_subject1
- benchmark:terminal-bench·BenchmarkTerminal-Bench
has_out_of_scope_reason2
- out-of-scope-reason:future-phase·OutOfScopeReasonFuture Phase
- out-of-scope-reason:implementation-detail·OutOfScopeReasonImplementation Detail
Incoming edges
None.