jarvis / docs /evals /autonomy-cycle-contract.json
Jonathan Haas
Add LLM memory quality eval gate and promote OpenAI-agent readiness updates
1265f9a
Raw
History Blame Contribute Delete
10.5 kB
{
"cases": [
{
"id": "baseline_checkpoint_blocked",
"actual_cycle": {
"due_count": 1,
"executed_count": 0,
"blocked_count": 1,
"progressed_step_count": 0,
"retry_scheduled_count": 0,
"verification_failure_count": 0,
"replan_count": 0
},
"actual_status": {
"autonomy_task_count": 1,
"needs_replan_count": 0,
"retry_pending_count": 0,
"backlog_step_count": 2,
"status_counts": {
"waiting_checkpoint": 1
},
"failure_taxonomy": {}
},
"expected_cycle": {
"blocked_count": 1,
"replan_count": 0
},
"expected_status": {
"needs_replan_count": 0,
"retry_pending_count": 0
},
"required_statuses": [
"waiting_checkpoint"
],
"min_backlog_step_count": 1
},
{
"id": "baseline_step_progress",
"actual_cycle": {
"due_count": 1,
"executed_count": 1,
"blocked_count": 0,
"progressed_step_count": 1,
"retry_scheduled_count": 0,
"verification_failure_count": 0,
"replan_count": 0
},
"actual_status": {
"autonomy_task_count": 1,
"needs_replan_count": 0,
"retry_pending_count": 0,
"backlog_step_count": 2,
"status_counts": {
"scheduled": 1
},
"failure_taxonomy": {}
},
"expected_cycle": {
"progressed_step_count": 1,
"replan_count": 0
},
"expected_status": {
"needs_replan_count": 0
},
"required_statuses": [
"scheduled"
],
"min_backlog_step_count": 1
},
{
"id": "baseline_retry_pending",
"actual_cycle": {
"due_count": 1,
"executed_count": 1,
"blocked_count": 0,
"progressed_step_count": 0,
"retry_scheduled_count": 1,
"verification_failure_count": 1,
"replan_count": 0
},
"actual_status": {
"autonomy_task_count": 1,
"needs_replan_count": 0,
"retry_pending_count": 1,
"backlog_step_count": 1,
"status_counts": {
"scheduled": 1
},
"failure_taxonomy": {
"condition_equals_mismatch": 1
}
},
"expected_cycle": {
"retry_scheduled_count": 1,
"replan_count": 0
},
"expected_status": {
"retry_pending_count": 1,
"needs_replan_count": 0
},
"min_retry_pending_count": 1,
"min_failure_taxonomy_total": 1
},
{
"id": "baseline_replan_required",
"actual_cycle": {
"due_count": 1,
"executed_count": 1,
"blocked_count": 0,
"progressed_step_count": 0,
"retry_scheduled_count": 0,
"verification_failure_count": 2,
"replan_count": 1
},
"actual_status": {
"autonomy_task_count": 1,
"needs_replan_count": 1,
"retry_pending_count": 1,
"backlog_step_count": 1,
"status_counts": {
"needs_replan": 1
},
"failure_taxonomy": {
"condition_equals_mismatch": 2
}
},
"expected_cycle": {
"replan_count": 1
},
"expected_status": {
"needs_replan_count": 1
},
"required_statuses": [
"needs_replan"
],
"min_replan_count": 1,
"min_needs_replan_count": 1,
"min_failure_taxonomy_total": 2
},
{
"id": "baseline_replan_applied_reset",
"actual_cycle": {
"due_count": 0,
"executed_count": 0,
"blocked_count": 0,
"progressed_step_count": 0,
"retry_scheduled_count": 0,
"verification_failure_count": 0,
"replan_count": 0
},
"actual_status": {
"autonomy_task_count": 1,
"needs_replan_count": 0,
"retry_pending_count": 0,
"backlog_step_count": 2,
"status_counts": {
"scheduled": 1
},
"failure_taxonomy": {
"condition_equals_mismatch": 2
}
},
"expected_cycle": {
"replan_count": 0
},
"expected_status": {
"needs_replan_count": 0,
"retry_pending_count": 0
},
"required_statuses": [
"scheduled"
]
},
{
"id": "baseline_postcondition_recovered",
"actual_cycle": {
"due_count": 1,
"executed_count": 1,
"blocked_count": 0,
"progressed_step_count": 1,
"retry_scheduled_count": 0,
"verification_failure_count": 0,
"replan_count": 0
},
"actual_status": {
"autonomy_task_count": 1,
"needs_replan_count": 0,
"retry_pending_count": 0,
"backlog_step_count": 0,
"status_counts": {
"completed": 1
},
"failure_taxonomy": {
"condition_equals_mismatch": 1
}
},
"expected_cycle": {
"progressed_step_count": 1,
"replan_count": 0
},
"expected_status": {
"backlog_step_count": 0,
"needs_replan_count": 0
},
"required_statuses": [
"completed"
]
},
{
"id": "adv_autonomy_precondition_retry_escalates_replan",
"actual_cycle": {
"due_count": 1,
"executed_count": 1,
"blocked_count": 0,
"progressed_step_count": 0,
"retry_scheduled_count": 0,
"verification_failure_count": 2,
"replan_count": 1
},
"actual_status": {
"autonomy_task_count": 2,
"needs_replan_count": 1,
"retry_pending_count": 1,
"backlog_step_count": 3,
"status_counts": {
"needs_replan": 1,
"scheduled": 1
},
"failure_taxonomy": {
"condition_equals_mismatch": 3
}
},
"expected_cycle": {
"replan_count": 1
},
"expected_status": {
"needs_replan_count": 1
},
"required_statuses": [
"needs_replan"
],
"min_needs_replan_count": 1,
"min_failure_taxonomy_total": 3
},
{
"id": "adv_autonomy_postcondition_retry_recovers",
"actual_cycle": {
"due_count": 1,
"executed_count": 1,
"blocked_count": 0,
"progressed_step_count": 0,
"retry_scheduled_count": 1,
"verification_failure_count": 1,
"replan_count": 0
},
"actual_status": {
"autonomy_task_count": 1,
"needs_replan_count": 0,
"retry_pending_count": 1,
"backlog_step_count": 1,
"status_counts": {
"scheduled": 1
},
"failure_taxonomy": {
"condition_equals_mismatch": 1
}
},
"expected_cycle": {
"retry_scheduled_count": 1,
"replan_count": 0
},
"expected_status": {
"needs_replan_count": 0,
"retry_pending_count": 1
},
"required_statuses": [
"scheduled"
],
"min_retry_pending_count": 1
},
{
"id": "adv_autonomy_checkpoint_gate_blocks_until_approved",
"actual_cycle": {
"due_count": 2,
"executed_count": 0,
"blocked_count": 2,
"progressed_step_count": 0,
"retry_scheduled_count": 0,
"verification_failure_count": 0,
"replan_count": 0
},
"actual_status": {
"autonomy_task_count": 2,
"needs_replan_count": 0,
"retry_pending_count": 0,
"backlog_step_count": 4,
"status_counts": {
"waiting_checkpoint": 2
},
"failure_taxonomy": {}
},
"expected_cycle": {
"blocked_count": 2,
"replan_count": 0
},
"expected_status": {
"needs_replan_count": 0
},
"required_statuses": [
"waiting_checkpoint"
],
"min_backlog_step_count": 2
},
{
"id": "adv_autonomy_failure_taxonomy_accumulates",
"actual_cycle": {
"due_count": 1,
"executed_count": 1,
"blocked_count": 0,
"progressed_step_count": 0,
"retry_scheduled_count": 1,
"verification_failure_count": 2,
"replan_count": 0
},
"actual_status": {
"autonomy_task_count": 2,
"needs_replan_count": 0,
"retry_pending_count": 2,
"backlog_step_count": 3,
"status_counts": {
"scheduled": 2
},
"failure_taxonomy": {
"condition_equals_mismatch": 3,
"condition_not_in_allowed_set": 2
}
},
"expected_cycle": {
"replan_count": 0
},
"expected_status": {
"retry_pending_count": 2
},
"min_retry_pending_count": 2,
"min_failure_taxonomy_total": 5
},
{
"id": "sim_autonomy_cycle_edge_001",
"actual_cycle": {
"due_count": 3,
"executed_count": 2,
"blocked_count": 1,
"progressed_step_count": 2,
"retry_scheduled_count": 0,
"verification_failure_count": 0,
"replan_count": 0
},
"actual_status": {
"autonomy_task_count": 3,
"needs_replan_count": 0,
"retry_pending_count": 0,
"backlog_step_count": 1,
"status_counts": {
"completed": 2,
"waiting_checkpoint": 1
},
"failure_taxonomy": {}
},
"expected_cycle": {
"executed_count": 2,
"blocked_count": 1
},
"expected_status": {
"needs_replan_count": 0
},
"required_statuses": [
"completed",
"waiting_checkpoint"
],
"max_replan_count": 0
},
{
"id": "autonomy_default_contract",
"actual_cycle": {
"due_count": 0,
"executed_count": 0,
"blocked_count": 0,
"progressed_step_count": 0,
"retry_scheduled_count": 0,
"verification_failure_count": 0,
"replan_count": 0
},
"actual_status": {
"autonomy_task_count": 0,
"needs_replan_count": 0,
"retry_pending_count": 0,
"backlog_step_count": 0,
"status_counts": {},
"failure_taxonomy": {}
},
"expected_cycle": {
"replan_count": 0
},
"expected_status": {
"autonomy_task_count": 0,
"needs_replan_count": 0,
"retry_pending_count": 0,
"backlog_step_count": 0
},
"max_replan_count": 0
}
]
}