ScaffoldBench

RunOpus 5.5repeat 2

Orchestrator-workers on Triage a new service desk ticket

claude-opus-5-5__orchestrator-workers__support-easy__r2

Score

1.000

passed
design ran as designed
Cost
$0.098
Latency
30.3s
setup 0.0s, grading 0.0s
Turns
12
Tool calls
8
Subagents
1
Tokens in / out
39k / 2.1k
29,396 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.
  • The task was decomposed: heldorchestrator-workers.decompositionThe orchestrator’s reply was read through its schema and held at least one subtask.
  • The fan-out happened: heldorchestrator-workers.fan-outEvery subtask got a worker, every worker ran, and the recorded subagent count matches.
  • The synthesis wrote the answer: heldorchestrator-workers.synthesisThe final answer came from the synthesis call, not from a worker’s report standing in for it.
  • Each call was shown the right tools: heldorchestrator-workers.toolsThe orchestrator and the synthesis were shown no tools and every worker the task’s full toolset.

Phases

Decompose: 0.0s to 6.6s, finished, 0 tool calls, $0.022
Worker 1: all: 6.6s to 23.1s, finished, 8 tool calls, $0.051
Synthesize: 23.1s to 30.3s, finished, 0 tool calls, $0.025

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

PhaseDurationCostTurnsToolsModel
Decompose6.6s$0.02220claude-opus-5-5
Worker 1: all16.5s$0.05198claude-opus-5-5
Synthesize7.2s$0.02510claude-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 orchestrator-workers --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
8
failed tool calls
0
distinct tools
7

Final answer

T-1041 ("VPN drops for the Berlin sales team") has been triaged. This is based on the worker's report; I had no way to check the system myself. The worker says it re-read the ticket afterwards and all three changes had saved.

  • Severity: sev3. The ticket is for the VPN and affects one team, and the escalation matrix grades that as sev3. The requester's seniority and the nine people affected don't change the grade, because the matrix decides severity.
  • Deadline: 2026-03-17T17:20:00Z. The ticket had no deadline. It was triaged at 09:20Z by the service desk clock, and sev3 has a 480-minute response target, so the deadline is 17:20Z.
  • Assigned to: E-301 Yuki Sato. The network-primary rotation owns the VPN. Yuki Sato is on call for that rotation when the deadline falls, on a shift from 08:00 to 20:00Z. The ticket's status changed from new to open.

No other ticket was changed, and nothing conflicted with the handbook.

The matrix gives sev3 no escalation rotation. If the outage spreads to a whole department or the company, the ticket's impact scope and severity would have to be raised before it could be escalated.

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.