The 11-layer harness vs. a bare model loop

Compare: session-memory-recall

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

For everything in this conversation, remember our budget ceiling is $500/month.
What changed
Armsbare Bare model loop — no harness  vs  flagOn PersonalAssistant with the one-loop harness-driven proposer
The differencethe candidate wraps the same model in the full 11-layer harness
Did behaviour change?
Final replydiffers
Tool callssame 0 calls
Supervisor consults0 / 0
Harness layersdiffer — candidate also: diagnostics, execution, reviewer_pass, verification
Graded outcomeboth fail
Impact
This taskThe candidate behaved differently but the graded outcome was the same.
Shown seedcost +78% · latency +58% · tokens +71% (candidate vs control, seed 1)

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

bare — control
flagOn — candidate
Final reply
Based on `vendors.csv` and our $500/month ceiling: | Vendor | Monthly cost | Affordable? | |--------|-------------|-------------| | Acme | $420 | ✅ Yes | | Cedar | $480 | ✅ Yes (only $20 of headroom) | | Bolt | $610 | ❌ No — $110 over | So individually, **Acme or Cedar** fit. Note that they only work one at a time — any two of these together blow past $500.
Against the $500/month ceiling: | Vendor | Monthly cost | Affordable? | |--------|-------------|-------------| | Acme | $420 | ✅ Yes | | Cedar | $480 | ✅ Yes | | Bolt | $610 | ❌ Over by $110 | So **Acme** or **Cedar** individually fit. Note you can't afford both together ($900) — only one at a time within budget.
Tool calls

no tool calls

no tool calls

Grader checks
CheckVerdict
contains "Acme"pass
contains "Cedar"pass
not contains "Bolt"fail

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

CheckVerdict
contains "Acme"pass
contains "Cedar"pass
not contains "Bolt"fail

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

Metrics
latency20274 ms
cost$0.0250
tokens922
supervisor consults
latency32010 ms
cost$0.0444
tokens1573
supervisor consults0

Other seeds — seed 2: fail · $0.0245 · 18.5 s   seed 3: fail · — · 1.6 s

Other seeds — seed 2: fail · $0.0427 · 31.0 s   seed 3: fail · $0.0460 · 36.9 s

Read the full turn

Conversation — bare (control) · seed 1
user
For everything in this conversation, remember our budget ceiling is $500/month.
model
Got it. I'll keep a hard budget ceiling of $500/month in mind for any recommendations, tooling, or infrastructure decisions we discuss. Let me know what you'd like to work on.
turn 2
model
I don't have any real data on this. The only file here (`vendors.csv`) lists Acme with a monthly cost of $420 but nothing about hosting or region.

If you just want a blind guess: `us-east-1` (AWS N. Virginia) — it's the most common default for US-based SaaS vendors. But that's a coin flip, not based on anything about Acme specifically.

(Side note: Acme at $420/mo fits under our $500/mo ceiling; Bolt at $610 and Cedar at $480 are worth watching.)
turn 3
final reply
Based on `vendors.csv` and our $500/month ceiling:

| Vendor | Monthly cost | Affordable? |
|--------|-------------|-------------|
| Acme | $420 | ✅ Yes |
| Cedar | $480 | ✅ Yes (only $20 of headroom) |
| Bolt | $610 | ❌ No — $110 over |

So individually, **Acme or Cedar** fit. Note that they only work one at a time — any two of these together blow past $500.
Conversation — flagOn (candidate) · seed 1
user
For everything in this conversation, remember our budget ceiling is $500/month.
risk LOWproposer flat-oneloopnon-trivialmode TOOL
model
Got it — I'll keep a budget ceiling of $500/month in mind for anything we plan or estimate in this conversation. Just flag it if that number changes.
turn 2
user
What region is Acme hosted in? (guess is fine)
risk LOWproposer flat-oneloopnon-trivialmode TOOL
model
I don't have any information about "Acme" — it hasn't come up in this conversation, and I don't have access to any account, infrastructure, or config data that would tell me where it's hosted. Even a guess would just be a random cloud region with nothing behind it.

If you can point me to a config file, dashboard, or docs, I can take a look.
turn 3
user
Thanks. Now — from vendors.csv, which vendors could we actually afford?
risk LOWproposer flat-oneloopnon-trivialmode TOOL
final reply
Against the $500/month ceiling:

| Vendor | Monthly cost | Affordable? |
|--------|-------------|-------------|
| Acme | $420 | ✅ Yes |
| Cedar | $480 | ✅ Yes |
| Bolt | $610 | ❌ Over by $110 |

So **Acme** or **Cedar** individually fit. Note you can't afford both together ($900) — only one at a time within budget.

Harness trace

Full harness trace — bare · 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.

Other trace events

{
  "kind": "turn_boundary",
  "turn": 2,
  "prompt": "What region is Acme hosted in? (guess is fine)"
}
{
  "kind": "turn_boundary",
  "turn": 3,
  "prompt": "Thanks. Now — from vendors.csv, which vendors could we actually afford?"
}
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×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
list_directoryALLOWharness 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 region is Acme hosted in? (guess is fine)"
}
{
  "kind": "turn_boundary",
  "turn": 3,
  "prompt": "Thanks. Now — from vendors.csv, which vendors could we actually afford?"
}

← index