Updated August 12, 2026

On this page

Check-execution types: ScriptExecution and AgentExecution (the two check kinds), CheckVerdict/CheckIssue/CriterionReview for structured agent verdicts, and CheckResult/CheckContext used by pre/post-check runners.

#protocols.src.orxtra.protocols._types._checks

#`protocols.src.orxtra.protocols._types._checks`

#`ScriptExecution`

ScriptExecution
Field	Type	Default
`callable`	`str`

#`AgentExecution`

AgentExecution
Field	Type	Default
`agent`	`str`
`task`	`str`
`block_threshold`	`Severity`
`variables`	`list[str]`	`field(default_factory=list)`

#`CheckVerdict`

CheckVerdict
Field	Type	Default
`verdict`	`str`
`issues`	`list[CheckIssue]`
`criteria_review`	`list[CriterionReview]`
`summary`	`str`

#`CheckIssue`

CheckIssue
Field	Type	Default
`severity`	`Severity`
`file`	`str	None`	`None`
`line_range`	`tuple[int, int]	None`	`None`
`description`	`str`
`blocking`	`bool`

#`CriterionReview`

CriterionReview
Field	Type	Default
`criterion`	`str`
`met`	`bool`
`evidence`	`str`

#`CheckResult`

CheckResult
Field	Type	Default
`passed`	`bool`
`message`	`str`
`details`	`dict[str, Any]	None`	`None`
`fix`	`Callable[..., Any]	None`	`None`

#`CheckContext`

CheckContext
Field	Type	Default
`variables`	`dict[str, Any]`
`agent_output`	`str	None`
`run_id`	`uuid.UUID`
`session_id`	`str	None`
`task_name`	`str`
`task_id`	`uuid.UUID`
`attempt`	`int`
`parent_task_id`	`uuid.UUID	None`

#`CheckAgentContext`

CheckAgentContext
Field	Type	Default
`task`	`str`
`agent_output`	`str`
`mechanical_results`	`str`
`task_name`	`str`
`attempt`	`int`
`notepad`	`str`