File size: 1,892 Bytes
b7e26de | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 | id: E4-deploy-002
title: "Public-Broadcast Playout Upgrade: Shell-State Persistence"
scenario:
domain: "Public-Service Broadcast Engineering"
primary_subscenario: "shell-state persistence during a split playout-system upgrade"
secondary_subscenarios:
- deterministic shell-state probes
- evidence-grounded overnight broadcast handoff
- cross-Harness command-execution comparison
harness:
dimension: E
subdimension: E4
subdimension_name: "shell-state persistence"
intensity:
level: L3
description: "Twelve ordered, setup-confirmed probes determine whether a split public-broadcast playout deployment can retain shell-local state, inherited process state, job control, and open resources across separate native command-execution calls."
input_mode: single_turn
tool_dependencies: [native_exec_v1, post_tool_v1, process_containment_v1, background_process_v1, business_cli_v1]
budget:
max_turns: 1
max_tool_calls: 160
max_wall_time_sec: 1800
expected_outputs:
- workspace/output/
verifier:
entrypoint: verifier/run.sh
scoring:
version: layered_100_v1
harness_score: verifier/harness_score.yaml
agent_score: verifier/agent_score.yaml
business_simulator:
version: unix_json_v1
scenario: simulator/scenario.json
server: simulator/server.py
cli: taskctl
execution_probe:
version: lifecycle_v1
requires: [native_exec_v1, post_tool_v1, process_containment_v1, background_process_v1, business_cli_v1]
execution:
mode: native
protocol: native_v1
profile: runtime/execution.json
rebuilt_from:
case: M-E3a-deploy-L3-001
what_changed: "Converted to one English input turn and a public-broadcast playout upgrade setting; every setup emits a success marker so setup failure cannot masquerade as state loss, and layered scoring directly measures cross-call recovery, trace integrity, evidence fidelity, and deterministic artifact correctness."
|