ScaffoldBench

RunOpus 5.5repeat 1

Gated-execute on Triage a new service desk ticket

claude-opus-5-5__gated-execute__support-easy__r1

Score

1.000

passed
design ran as designed
Cost
$0.141
Latency
35.3s
setup 0.0s, grading 0.0s
Turns
24
Tool calls
17
Subagents
3
Tokens in / out
61k / 2.3k
45,581 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 35.3s, finished, 7 tool calls, $0.043
Review 1: set_severity: 6.5s to 14.3s, finished, 2 tool calls, $0.041
Review 2: extend_sla: 14.3s to 20.7s, finished, 3 tool calls, $0.025
Review 3: assign_ticket: 22.5s to 31.1s, finished, 5 tool calls, $0.033

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

PhaseDurationCostTurnsToolsModel
Execute35.3s$0.04387claude-opus-5-5
Review 1: set_severity7.8s$0.04142claude-opus-5-5
Review 2: extend_sla6.4s$0.02553claude-opus-5-5
Review 3: assign_ticket8.6s$0.03375claude-opus-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-easy --models claude-opus-5-5 --repeats 1

Score breakdown

  • passSeverity matches the escalation matrix for the service and impact scope
    1.00weight 0.27

    severity=sev3 (matrix says sev3)

  • passSLA deadline is the triage time plus the response target for that severity
    1.00weight 0.27

    sla_due_at=2026-03-17T17:20:00Z (expected 2026-03-17T17:20:00Z)

  • passAssigned to the engineer on call for the owning rotation at the deadline
    1.00weight 0.27

    assignee=E-301 (expected E-301)

  • passTicket left in status open
    1.00weight 0.09
  • passEscalation matrix read before the severity was set
    1.00weight 0.09

Penalties

No penalties were applied.

Scorer metadata

tool calls
17
failed tool calls
0
distinct tools
8

Final answer

I've triaged T-1041 ("VPN drops for the Berlin sales team"). It's now sev3, due 2026-03-17T17:20:00Z, and assigned to E-301 Yuki Sato with status open.

  • Severity: sev3. The escalation matrix rates a VPN problem affecting one team as sev3. About nine people on the Berlin sales floor can't connect, so the impact is team-level.
  • Deadline: 17:20 UTC. The ticket had no deadline, so I set one at triage time (09:20 UTC on the service desk clock) plus sev3's 480-minute response target.
  • Assignee: E-301 Yuki Sato. The network-primary rotation owns VPN, and Yuki is on call for it at the deadline (shift runs 08:00–20:00 UTC).

I didn't change any other ticket, and nothing I found conflicted with the task or the handbook rules.

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.