{"id":"task-pack-handoff-envelope-smoke","prompt":"Compile a settled source plan into an executable task pack handoff.","baseline_output":"Here is a list of tasks to do next:\n1. Update the validator.\n2. Add tests.\n3. Run the suite.","with_skill_output":"decision: compile\nreason_code: task_pack_worthwhile\nTask Pack Contract: present\nsource_plan_hash: sha256:0123456789abcdef0123456789abcdef0123456789abcdef0123456789abcdef\ndeterministic_handoff: true\nvalidity_scope: identity-freshness-structure-only\nnext_action: spec-work-task-pack","assertions":[{"id":"decision-envelope","description":"The output includes the final decision envelope fields needed for deterministic handoff.","required":["decision: compile","reason_code:","deterministic_handoff: true","next_action: spec-work-task-pack"],"forbidden":["approval state"],"weight":2,"failure_type":"missing_decision_envelope"},{"id":"task-pack-contract","description":"The output names the machine-readable task pack contract and source plan hash.","required":["Task Pack Contract","source_plan_hash: sha256:"],"forbidden":["pending-tooling"],"weight":2,"failure_type":"missing_task_pack_contract"}],"metadata":{"case_type":"yao-smoke","evidence_status":"recorded_static_smoke","authoritative_runner":"scripts/spec-write-tasks/run-output-evals.js"},"human_review":{"expected_winner":"with_skill"}}
{"id":"high-risk-review-handoff-smoke","prompt":"Compile a task pack with high-risk review gates without auto-dispatching downstream review.","baseline_output":"The task pack is ready, so I will run the next review workflow automatically.","with_skill_output":"decision: compile\nnext_action: review-task-pack\nreason_code: high_risk_review_gate\ndispatch_authorization: missing\ncopy_ready_review_invocation: current-host document-review invocation\nsemantic_posture: generated-this-run","assertions":[{"id":"bounded-review-handoff","description":"High-risk task packs return review handoff metadata without claiming downstream dispatch authorization.","required":["next_action: review-task-pack","dispatch_authorization: missing","copy_ready_review_invocation"],"forbidden":["run the next review workflow automatically"],"weight":2,"failure_type":"unsafe_auto_dispatch"},{"id":"semantic-posture","description":"The output identifies semantic posture instead of treating review_gate as approval state.","required":["semantic_posture:"],"forbidden":["approved"],"weight":1,"failure_type":"missing_semantic_posture"}],"metadata":{"case_type":"yao-smoke","evidence_status":"recorded_static_smoke","authoritative_runner":"scripts/spec-write-tasks/run-output-evals.js"},"human_review":{"expected_winner":"with_skill"}}