Trajectory Supervisor (mid-task stalls)

Compare: sup-mid-control-clean-a

Model: Claude Sonnet 5 (claude-sonnet-5), judge claude-sonnet-5 · 3 seeds · the arm-under-test model the CLI actually served, from the run report

Prompt

What is the default region in config/app.yaml?
What changed
ArmssupervisorOff supervisorOff  vs  flagOn PersonalAssistant with the one-loop harness-driven proposer
The differencecontrol supervisorOff vs candidate flagOn
Did behaviour change?
Final replyidentical
Tool callssame 0 calls
Supervisor consults0 / 0
Harness layerssame set fired
Graded outcome3/3 vs 3/3 pass
Impact
This taskNo behavioural change on this task — same reply, same tool calls, same harness layers. The candidate did the extra work for an identical result.
Shown seedcost -1% · latency -2% · tokens -1% (candidate vs control, seed 1)

Side by side — control (left) vs candidate (right), seed 1

supervisorOff — control
flagOn — candidate
Final reply
ap-south-9
ap-south-9
Tool calls

no tool calls

no tool calls

Grader checks
CheckVerdict
contains "ap-south-9"pass

success yes  ·  hallucination no  ·  unauthorized effect no  ·  recovered n/a

CheckVerdict
contains "ap-south-9"pass

success yes  ·  hallucination no  ·  unauthorized effect no  ·  recovered n/a

Metrics
latency17259 ms
cost$0.0118
tokens681
supervisor consults0
latency16966 ms
cost$0.0117
tokens671
supervisor consults0

Other seeds — seed 2: pass · $0.0116 · 16.8 s   seed 3: pass · $0.0116 · 17.1 s

Other seeds — seed 2: pass · $0.0118 · 16.4 s   seed 3: pass · $0.0116 · 17.2 s

Read the full turn

Conversation — supervisorOff (control) · seed 1
user
What is the default region in config/app.yaml?
risk LOWproposer flat-oneloopnon-trivialmode TOOL
model
The default region in `config/app.yaml` is `ap-south-9`.
turn 2
user
What is the log level?
risk LOWproposer flat-oneloopnon-trivialmode TOOL
model
The log level in `config/app.yaml` is `warn`.
turn 3
user
And which region did you say was the default? Answer in one word.
risk LOWproposer flat-oneloopnon-trivialmode TOOL
final reply
ap-south-9
Conversation — flagOn (candidate) · seed 1
user
What is the default region in config/app.yaml?
risk LOW
model
The default region in `config/app.yaml` is `ap-south-9`.
turn 2
proposer posthoctrivialmode FAST
user
What is the log level?
risk LOWproposer flat-oneloopnon-trivialmode TOOL
model
The log level in `config/app.yaml` is `warn`.
turn 3
user
And which region did you say was the default? Answer in one word.
risk LOWproposer flat-oneloopnon-trivialmode TOOL
final reply
ap-south-9

Harness trace

Full harness trace — supervisorOff · seed 1

The harness runs on every turn. Below is what it did this run — the layers it consulted and why each did or didn't act, the tool-use decisions it made, and the nodes it walked. Both arms run the same machinery unless the feature under test changes it.

risk LOWproposer flat-oneloopmode TOOLplan none

Harness layers

LayerActed?Why
world_model— ×3single LOW-risk task, no durable fact stated — observation only
evidence_reasoning— ×3single low-stakes observation is sufficient
hypothesis— ×6single clear LOW-risk task — no competing explanation worth surfacing
contradiction— ×6fewer than 2 beliefs — nothing to compare
diagnosticsacted ×6Health: nominal
control_state— ×6NORMAL
planning— ×3one eligible task — serial execution
executionacted ×3module_type=business_logic
verificationacted ×3all applicable layers passed
recovery— ×3task completed — nothing to recover from
reviewer_passacted ×3Success criterion not covered by any belief: "Respond helpfully, accurately, and safely to the user request."

Tool-policy decisions

ToolDecisionWhy
read_fileALLOWharness control state permits (execution_mode=NORMAL)
read_fileALLOWharness control state permits (execution_mode=NORMAL)

Node path

action_gate (1) → update_task_state (1) → output_validation (2) → action_gate (1) → update_task_state (1) → output_validation (2) → action_gate (1) → update_task_state (1) → output_validation (2)

Other trace events

{
  "kind": "turn_boundary",
  "turn": 2,
  "prompt": "What is the log level?"
}
{
  "kind": "turn_boundary",
  "turn": 3,
  "prompt": "And which region did you say was the default? Answer in one word."
}
Full harness trace — flagOn · seed 1

The harness runs on every turn. Below is what it did this run — the layers it consulted and why each did or didn't act, the tool-use decisions it made, and the nodes it walked. Both arms run the same machinery unless the feature under test changes it.

risk LOWproposer flat-oneloopmode TOOLplan none

Harness layers

LayerActed?Why
world_model— ×2single LOW-risk task, no durable fact stated — observation only
evidence_reasoning— ×2single low-stakes observation is sufficient
hypothesis— ×4single clear LOW-risk task — no competing explanation worth surfacing
contradiction— ×4fewer than 2 beliefs — nothing to compare
diagnosticsacted ×4Health: nominal
control_state— ×4NORMAL
planning— ×2one eligible task — serial execution
executionacted ×2module_type=business_logic
verificationacted ×2all applicable layers passed
recovery— ×2task completed — nothing to recover from
reviewer_passacted ×2Success criterion not covered by any belief: "Respond helpfully, accurately, and safely to the user request."

Tool-policy decisions

ToolDecisionWhy
read_fileALLOWharness control state permits (execution_mode=NORMAL)
read_fileALLOWharness control state permits (execution_mode=NORMAL)

Node path

action_gate (1) → update_task_state (1) → output_validation (2) → action_gate (1) → update_task_state (1) → output_validation (2)

Other trace events

{
  "kind": "turn_boundary",
  "turn": 2,
  "prompt": "What is the log level?"
}
{
  "kind": "turn_boundary",
  "turn": 3,
  "prompt": "And which region did you say was the default? Answer in one word."
}

← index