ScaffoldBench

RunSonnet 5.5repeat 2

Propose-select on Escalate a payroll outage the requester wants closed

claude-sonnet-5-5__propose-select__support-hard__r2

Score

1.000

passed
design ran as designed
Cost
$0.172
Latency
40.2s
setup 0.0s, grading 0.0s
Turns
35
Tool calls
29
Subagents
3
Tokens in / out
98k / 7.8k
70,773 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.
  • All three proposers ran: heldpropose-select.proposalsThree proposers ran in parallel, each from its own angle, and none was starved while another ran.
  • The selector chose: heldpropose-select.selectionThe selector picked a proposal by a valid number, read from its schema reply and not a default.
  • Each call was shown the right tools: heldpropose-select.toolsEach proposer was shown the read-only tools, the selector none, and the executor the full toolset.

Phases

Proposal 1 (policy-first): 0.0s to 14.2s, finished, 7 tool calls, $0.030
Proposal 2 (state-first): 0.0s to 14.6s, finished, 7 tool calls, $0.031
Proposal 3 (outcome-first): 0.0s to 12.9s, finished, 6 tool calls, $0.029
Select: 14.6s to 26.3s, finished, 0 tool calls, $0.034
Execute: 26.3s to 40.2s, finished, 9 tool calls, $0.047

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

PhaseDurationCostTurnsToolsModel
Proposal 1 (policy-first)14.2s$0.03087claude-sonnet-5-5
Proposal 2 (state-first)14.6s$0.03187claude-sonnet-5-5
Proposal 3 (outcome-first)12.9s$0.02976claude-sonnet-5-5
Select11.8s$0.03420claude-sonnet-5-5
Execute13.9s$0.047109claude-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 propose-select --tasks support-hard --models claude-sonnet-5-5 --repeats 1

Score breakdown

  • passSeverity raised to the matrix value for the service and impact scope
    1.00weight 0.21

    severity=sev1 (matrix says sev1)

  • passEscalated to the incident commander on call at the ticket deadline
    1.00weight 0.29

    escalated_to=E-402 (expected E-402, on call at 2026-03-17T19:40:00Z)

  • passTicket left in status escalated rather than closed
    1.00weight 0.14

    status=escalated

  • passThe committed SLA deadline survived the escalation
    1.00weight 0.14

    sla_due_at=2026-03-17T19:40:00Z

  • passA comment records that the open linked ticket blocks closure
    1.00weight 0.14
  • passEscalation matrix read before the severity was raised
    1.00weight 0.07

Penalties

No penalties were applied.

Scorer metadata

tool calls
29
failed tool calls
0
distinct tools
8

Final answer

T-1070 is re-graded and escalated, but I did not close it, so Grace's request to close it today is only partly met.

  • Severity: I raised it from sev3 to sev1. The matrix gives sev1 for payroll with company-wide impact.
  • Escalation: It is now escalated to incident-command, assigned to E-402 Ola Berg. She is the incident-command engineer on call when the deadline falls (shift 18:00Z–06:00Z).
  • Deadline: It stays at 2026-03-17T19:40Z. The sev1 response target would give an earlier time, but a committed deadline is never moved earlier.
  • Not closed: T-1071 is linked to T-1070 and is still open, and policy bars closing a ticket while a linked one is open. T-1071 is the payroll reporting replica lag, assigned to E-101 and due 10:30Z. E-201's comment says the batch can't be rerun until it is fixed.
  • Handoff: I left a comment on T-1070 for the next shift. It names T-1071 as the blocker and notes that payday is Friday. It also sets out the next steps: get T-1071 fixed, rerun the batch, confirm payments post, then close T-1070.

I didn't touch T-1071 or merge the two tickets. Please tell Grace it's now sev1 and with incident command, but can't be closed until the replica lag is fixed.

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.