ScaffoldBench

RunSonnet 5.5repeat 1

Gated-execute on Run the major-incident runbook for a CI outage

claude-sonnet-5-5__gated-execute__support-major-incident__r1

Score

0.820

wrong answer
design ran as designed
Cost
$0.510
Latency
2m 43s
setup 0.0s, grading 0.0s
Turns
163
Tool calls
130
Subagents
16
Tokens in / out
476k / 16k
388,738 read from cache

Did the design run?

every check held
  • Run completed: heldharness.run-completedThe environment was built, the architecture returned and grading finished, none of them throwing.
  • No API failures: heldharness.apiNo model call failed at the API.
  • Every call ended: heldharness.calls-endedEvery call ended with a result, or was stopped by the run's own limits.
  • No harness tool denied: heldharness.harness-toolsThe sandbox guard never denied a harness tool, nor a tool the session itself offered.
  • Tool servers connected: heldharness.mcp-connectedEvery tool server the harness gave a call was connected when the session started.
  • No foreign tool servers: heldharness.mcp-hermeticNo call loaded a tool server the harness did not give it.
  • Builtin tools: heldharness.builtin-toolsA check recorded with the run.
  • Structured replies arrived: heldharness.structured-outputEvery call that asked for structured output and ran to the end got it.
  • No silent fallbacks: heldharness.not-degradedNo phase fell back on a default (an empty plan, a default winner) instead of its output.
  • Every state change was reviewed: heldgated-execute.reviewedEvery call that changed the world had been allowed by a review before it ran, counted from the world’s own journal.
  • Every decision was read: heldgated-execute.decisionsEvery review that ran to its end gave its decision through the schema; an unreadable one is a denial.
  • Each call was shown the right tools: heldgated-execute.toolsThe executor was shown the task’s full toolset and each reviewer the read-only tools.

Phases

Execute: 0.0s to 2m 43s, finished, 38 tool calls, $0.108
Review 1: create_incident: 12.9s to 19.0s, finished, 3 tool calls, $0.019
Review 2: page_engineer: 21.3s to 27.6s, finished, 5 tool calls, $0.017
Review 3: assign_incident_role: 30.2s to 40.0s, finished, 5 tool calls, $0.019
Review 4: link_to_incident: 40.0s to 50.8s, finished, 11 tool calls, $0.039
Review 5: set_severity: 54.0s to 1m 1s, finished, 6 tool calls, $0.023
Review 6: set_severity: 1m 1s to 1m 9s, finished, 6 tool calls, $0.025
Review 7: set_severity: 1m 9s to 1m 16s, finished, 6 tool calls, $0.025
Review 8: set_severity: 1m 16s to 1m 24s, finished, 6 tool calls, $0.026
Review 9: set_severity: 1m 24s to 1m 31s, finished, 4 tool calls, $0.023
Review 10: set_severity: 1m 31s to 1m 39s, finished, 5 tool calls, $0.022
Review 11: set_severity: 1m 39s to 1m 46s, finished, 5 tool calls, $0.025
Review 12: assign_incident_role: 1m 46s to 1m 54s, finished, 10 tool calls, $0.031
Review 13: post_status_update: 1m 58s to 2m 7s, finished, 4 tool calls, $0.025
Review 14: extend_sla: 2m 7s to 2m 18s, finished, 7 tool calls, $0.029
Review 15: extend_sla: 2m 18s to 2m 28s, finished, 5 tool calls, $0.028
Review 16: create_ticket: 2m 28s to 2m 36s, finished, 4 tool calls, $0.025

Drawn from the phase records: timing and cost per call, no tool calls or messages.

PhaseDurationCostTurnsToolsModel
Execute2m 43s$0.1083938claude-sonnet-5-5
Review 1: create_incident6.2s$0.01953claude-sonnet-5-5
Review 2: page_engineer6.3s$0.01775claude-sonnet-5-5
Review 3: assign_incident_role9.8s$0.01975claude-sonnet-5-5
Review 4: link_to_incident10.8s$0.0391311claude-sonnet-5-5
Review 5: set_severity7.0s$0.02386claude-sonnet-5-5
Review 6: set_severity7.9s$0.02586claude-sonnet-5-5
Review 7: set_severity7.5s$0.02586claude-sonnet-5-5
Review 8: set_severity7.5s$0.02686claude-sonnet-5-5
Review 9: set_severity7.5s$0.02364claude-sonnet-5-5
Review 10: set_severity7.2s$0.02275claude-sonnet-5-5
Review 11: set_severity7.5s$0.02575claude-sonnet-5-5
Review 12: assign_incident_role8.3s$0.0311210claude-sonnet-5-5
Review 13: post_status_update8.3s$0.02564claude-sonnet-5-5
Review 14: extend_sla10.9s$0.02997claude-sonnet-5-5
Review 15: extend_sla10.8s$0.02875claude-sonnet-5-5
Review 16: create_ticket7.7s$0.02564claude-sonnet-5-5

The call-by-call replay is not published for this run. Traces are large, so the repository keeps 9 of this run set’s 900 and leaves the rest out; the score, phases and final answer on this page come from the committed run record.

Runs you can replay

Record this cell yourself: npm run bench -- run --strategies gated-execute --tasks support-major-incident --models claude-sonnet-5-5 --repeats 1

Score breakdown

  • passAn incident is open for the build pipeline
    1.00weight 0.04
  • passThe incident-command engineer on call now commands the incident
    1.00weight 0.08

    commander=E-401 (expected E-401)

  • passThe commander was paged
    1.00weight 0.04
  • passT-3101 linked to the incident
    1.00weight 0.04
  • passT-3101 kept open, raised to sev1 with its deadline intact
    1.00weight 0.02

    status=open severity=sev1 due=2026-03-30T18:52:00Z (expected sev1, 2026-03-30T18:52:00Z)

  • passT-3102 linked to the incident
    1.00weight 0.04
  • passT-3102 kept open, raised to sev1 with its deadline intact
    1.00weight 0.02

    status=open severity=sev1 due=2026-03-30T18:58:00Z (expected sev1, 2026-03-30T18:58:00Z)

  • passT-3103 linked to the incident
    1.00weight 0.04
  • passT-3103 kept open, raised to sev1 with its deadline intact
    1.00weight 0.02

    status=open severity=sev1 due=2026-03-30T19:01:00Z (expected sev1, 2026-03-30T19:01:00Z)

  • passT-3104 linked to the incident
    1.00weight 0.04
  • passT-3104 kept open, raised to sev1 with its deadline intact
    1.00weight 0.02

    status=open severity=sev1 due=2026-03-31T15:05:00Z (expected sev1, 2026-03-31T15:05:00Z)

  • passT-3105 linked to the incident
    1.00weight 0.04
  • passT-3105 kept open, raised to sev1 with its deadline intact
    1.00weight 0.02

    status=open severity=sev1 due=2026-04-01T15:07:00Z (expected sev1, 2026-04-01T15:07:00Z)

  • passT-3106 linked to the incident
    1.00weight 0.04
  • passT-3106 kept open, raised to sev1 with its deadline intact
    1.00weight 0.02

    status=new severity=sev1 due=2026-03-30T16:20:00Z (expected sev1, 2026-03-30T16:20:00Z)

  • passT-3109 linked to the incident
    1.00weight 0.04
  • passT-3109 kept open, raised to sev1 with its deadline intact
    1.00weight 0.02

    status=new severity=sev1 due=2026-03-30T16:20:00Z (expected sev1, 2026-03-30T16:20:00Z)

  • failT-3112 linked to the incident
    0.00weight 0.04
  • failT-3112 kept open, raised to sev1 with its deadline intact
    0.00weight 0.02

    status=new severity=unset due=- (expected sev1, 2026-03-30T16:20:00Z)

  • failT-3113 linked to the incident
    0.00weight 0.04
  • failT-3113 kept open, raised to sev1 with its deadline intact
    0.00weight 0.02

    status=open severity=sev3 due=2026-03-31T15:13:00Z (expected sev1, 2026-03-31T15:13:00Z)

  • failT-3114 linked to the incident
    0.00weight 0.04
  • failT-3114 kept open, raised to sev1 with its deadline intact
    0.00weight 0.02

    status=open severity=sev3 due=2026-03-31T15:16:00Z (expected sev1, 2026-03-31T15:16:00Z)

  • passThe engineer covering comms right now is the communications lead
    1.00weight 0.08

    comms=E-803 (expected E-803)

  • passA status update in the runbook form, with the next update due within 30 minutes
    1.00weight 0.08
  • passA postmortem ticket for the service, linked to the incident, due at close of business five business days on
    1.00weight 0.08

    T-3115 due=2026-04-08T17:00:00Z

Penalties

No penalties were applied.

Scorer metadata

tool calls
130
failed tool calls
1
distinct tools
19

Final answer

INC-1 is open for the company-wide build-pipeline outage, and I ran it through the current runbook, KB-40. I ignored KB-39 because KB-40 supersedes it.

  • Incident and commander: I opened INC-1 first. I then paged E-401 Priya Raman, the incident-command engineer on call now, and only her, and made her commander. The page can't be recalled.
  • Linked reports: T-3101, T-3102, T-3103, T-3104, T-3105, T-3106 and T-3109 are linked to INC-1. All seven describe the same no-runners outage.
  • Left out: I left out T-3107 (a flaky test) and T-3108 (a local Xcode link failure) because they are separate problems.
  • Severity: All seven linked reports are now sev1. Deadlines already committed stay as they were. T-3106 and T-3109 had none, so I gave each 16:20 UTC, which is the triage time plus the 60-minute sev1 target.
  • Comms lead: The comms rotation's E-801 Mia Laurent is on sick leave today. Under the leave rule the backup rotation, comms-secondary, covers, so I named E-803 Elif Demir, who is on call there now.
  • Status update: I posted the first update at 15:20 UTC. It says "Impact: All company build pipelines are affected… Started: 14:52 UTC. Next update: 15:45 UTC." The 14:52 start is when the earliest report, T-3101, was created.
  • Postmortem: T-3115 is linked to INC-1 and due 2026-04-08 at 17:00 UTC. That is the 5th business day after today, counting Good Friday (Apr 3) and Easter Monday (Apr 6) as holidays.

The next status update is due by 15:45 UTC.

Provenance of the numbers on this page

run set
scaffold-or-scale-v1
models
Haiku 4.5, Sonnet 5.5, Opus 5.5
effort
high
repeats
2
fixtures
replay
runs
900 / 900
strategy cost
$354.44
judge cost
$9.19
wall clock
9h 48m
git
fce5807b2d

Cost excludes judge calls, which are tracked separately so a scoring model never flatters or penalises an architecture. Every figure on this site is read from results/ at build time.