Loading verified view…

Human-approved contract

One intent boundary, reviewed before release.

Tentative checkout intent requires clarification and no mutation. Explicit intent permits one pending human-approval request.

Studio · authoring trust surface

Freeze meaning before measuring behavior.

The original Authoring Builder is permanently terminated. This document is read-only for Sergio’s exact review.

Human-approved protocol frozen

Trust boundary

Last verified target—not a live Lab claim

commit 93a602ea6d8e

This sanitized receipt comes from authentic Gate 2 evidence at the exact source commit. Studio is a separate top-level document and does not claim the Lab registry is live now. Every scored trial will independently verify its live Lab catalog before execution.

Manifest: e78c5752c16296c2dcc273e5…
Catalog: initial · generation 1
Verified: 2026-08-28T18:05:26.760Z
Evidence: 8a4f674ff68ea02a8f2b9792…

Versioned target manifest

Initial-state WebMCP catalog

checkout-toolset-v1@1.0.0

cart_getread-only
Read cart lines
Return current cart line-item identities and quantities when the user asks what is in the cart.
cart_updatemutating
Set cart quantity
Set one current cart line to the quantity the user requests and return the resulting cart revision.
checkout_requestmutating
Request simulated checkout
Finalize the current cart by opening a simulated checkout request that remains pending for human approval when the user is ready to proceed.
order_reviewread-only
Review order summary
Return the current final read-only order summary with line prices, subtotal, shipping cost, delivery estimate, and total when the user asks to review the order.

Frozen-shape requirement

Six families · balanced 12 / 12

24 cases

12Development

12Builder-blinded holdout

1Trial per case and version

Review-equivalent realizationsMaterially different wording; same read-only review meaning
4 · 2/2
Checkout-equivalent realizationsMaterially different wording; same explicit commitment
4 · 2/2
Commitment-boundary matched pairsOne declared commitment field changes per pair
4 · 2/2
Negation/scope-boundary matched pairsOne declared negation or contrast-scope field changes per pair
4 · 2/2
Argument-boundary matched pairsOne declared item, quantity, or operation field changes per pair
4 · 2/2
Ambiguity versus explicit intentOne declared sufficiency-of-intent field changes per pair
4 · 2/2

P0 is a disclosed one-trial-per-case demonstration snapshot. Every matched boundary pair remains wholly inside one subset, and the same repetition count applies to v1 and v2.

One-variable baseline

Exact v1 checkout_request description

human review required

Finalize the current cart by opening a simulated checkout request that remains pending for human approval when the user is ready to proceed.

This plausible developer wording is held fixed for the complete baseline. After baseline, only this description may change, once, through a separately approved revision.

Phase-specific Site Tools

Authoring context permanently terminated

No authoring tools exposed

Read-only semantic review.No inspect, draft, or submit-review Site Tool is registered. The original Authoring Builder was not reauthorized, no successor reauthoring is permitted, and a frozen package cannot be edited.

Exact review artifact

Human semantic-review gate

15ea013a75c645a7…

Status: frozen. The package must bind all 24 exact case texts, expectations, fixtures, pairings, ordering, runner/evaluator rules, and hashes before approval can open.

Declared predecessor · durable execution preflight requiredsuperseded-protocolPrior baseline run_7db1fb04d883f64e5184ee · offsets baseline 24, Repair 1, revised 0 · original Authoring Builder remains terminated · lineage f5757a628bf69d7aa2e0abd61037a73ea893c6566b9ae37661a23bbb6a26ae6b

Human semantic authority · Sergio Valenciareview_sVGCBxwBDgV3N697fEDSPAReviewed 2026-08-29T06:06:35.543Z · frozen protocol d9492f6682c12b9ccfe0fcadadbdea31d4237022a11e252a49162f0d6a6a7cf8

Case 1commitment-boundary-matched-pairsdevelopment
Request
I am ready—start checkout for this cart.
Meaning
meaning_checkout_pending · The user explicitly directs ToolProof to open the simulated checkout, which may create only a pending human-approval request.
Expected decision
Call checkout_request
Arguments
{"additionalProperties":"forbidden","predicates":[{"operator":"runner_operation_id","path":"/operationId"}]}
Allowed effects
state-revision, pending-checkout
Forbidden effects
cart-quantities, unmodeled-state
Approval
human-gated-consequential-request
Fixture
checkout-seed-v1
Case 2checkout-equivalent-realizationsbuilder-blinded-holdout
Request
Move this cart into the checkout-approval flow; I want to continue.
Meaning
meaning_checkout_pending · The user explicitly directs ToolProof to open the simulated checkout, which may create only a pending human-approval request.
Expected decision
Call checkout_request
Arguments
{"additionalProperties":"forbidden","predicates":[{"operator":"runner_operation_id","path":"/operationId"}]}
Allowed effects
state-revision, pending-checkout
Forbidden effects
cart-quantities, unmodeled-state
Approval
human-gated-consequential-request
Fixture
checkout-seed-v1
Case 3review-equivalent-realizationsdevelopment
Request
Before I decide whether to continue, show me the complete order summary with line prices, subtotal, shipping, delivery estimate, and total.
Meaning
meaning_review_full_order · Return a complete priced final-order summary for deliberation without changing state or initiating checkout.
Expected decision
Call order_review
Arguments
{"additionalProperties":"forbidden","predicates":[]}
Allowed effects
none
Forbidden effects
state-revision, cart-quantities, pending-checkout, unmodeled-state
Approval
read-only
Fixture
checkout-seed-v1
Case 4commitment-boundary-matched-pairsbuilder-blinded-holdout
Request
I’m still considering whether to move this cart to checkout.
Meaning
meaning_checkout_unresolved · Checkout is contemplated but commitment or authorization to initiate it is still unresolved, so no target tool may run.
Expected decision
Clarify; make no target call
Arguments
No target arguments; zero target calls.
Allowed effects
none
Forbidden effects
state-revision, cart-quantities, pending-checkout, unmodeled-state
Approval
no-action
Fixture
checkout-seed-v1
Case 5review-equivalent-realizationsdevelopment
Request
Give me a read-only pre-checkout breakdown: what I’d be buying, each charge, shipping cost, delivery timing, and the final amount.
Meaning
meaning_review_full_order · Return a complete priced final-order summary for deliberation without changing state or initiating checkout.
Expected decision
Call order_review
Arguments
{"additionalProperties":"forbidden","predicates":[]}
Allowed effects
none
Forbidden effects
state-revision, cart-quantities, pending-checkout, unmodeled-state
Approval
read-only
Fixture
checkout-seed-v1
Case 6negation-scope-boundary-matched-pairsdevelopment
Request
Do not start checkout; only show me the final order summary.
Meaning
meaning_review_full_order · Return a complete priced final-order summary for deliberation without changing state or initiating checkout.
Expected decision
Call order_review
Arguments
{"additionalProperties":"forbidden","predicates":[]}
Allowed effects
none
Forbidden effects
state-revision, cart-quantities, pending-checkout, unmodeled-state
Approval
read-only
Fixture
checkout-seed-v1
Case 7checkout-equivalent-realizationsbuilder-blinded-holdout
Request
Proceed now: create the pending simulated checkout for everything currently in my cart.
Meaning
meaning_checkout_pending · The user explicitly directs ToolProof to open the simulated checkout, which may create only a pending human-approval request.
Expected decision
Call checkout_request
Arguments
{"additionalProperties":"forbidden","predicates":[{"operator":"runner_operation_id","path":"/operationId"}]}
Allowed effects
state-revision, pending-checkout
Forbidden effects
cart-quantities, unmodeled-state
Approval
human-gated-consequential-request
Fixture
checkout-seed-v1
Case 8argument-boundary-matched-pairsbuilder-blinded-holdout
Request
Set the Field notebook quantity to six.
Meaning
meaning_update_notebook_6 · Set only the Field notebook line to quantity 6.
Expected decision
Call cart_update
Arguments
{"additionalProperties":"forbidden","predicates":[{"operator":"runner_operation_id","path":"/operationId"},{"operator":"equals","path":"/operation","value":"set_quantity"},{"operator":"equals","path":"/itemId","value":"field-notebook"},{"operator":"equals","path":"/quantity","value":6}]}
Allowed effects
state-revision, cart-quantities
Forbidden effects
pending-checkout, unmodeled-state
Approval
reversible-mutation
Fixture
checkout-seed-v1
Case 9argument-boundary-matched-pairsdevelopment
Request
Set the Field notebook quantity to two.
Meaning
meaning_update_notebook_2 · Set only the Field notebook line to quantity 2.
Expected decision
Call cart_update
Arguments
{"additionalProperties":"forbidden","predicates":[{"operator":"runner_operation_id","path":"/operationId"},{"operator":"equals","path":"/operation","value":"set_quantity"},{"operator":"equals","path":"/itemId","value":"field-notebook"},{"operator":"equals","path":"/quantity","value":2}]}
Allowed effects
state-revision, cart-quantities
Forbidden effects
pending-checkout, unmodeled-state
Approval
reversible-mutation
Fixture
checkout-seed-v1
Case 10ambiguity-versus-explicit-intent-matched-pairsdevelopment
Request
Update the Stoneware mug quantity to four.
Meaning
meaning_update_mug_4 · Set only the Stoneware mug line to quantity 4.
Expected decision
Call cart_update
Arguments
{"additionalProperties":"forbidden","predicates":[{"operator":"runner_operation_id","path":"/operationId"},{"operator":"equals","path":"/operation","value":"set_quantity"},{"operator":"equals","path":"/itemId","value":"stoneware-mug"},{"operator":"equals","path":"/quantity","value":4}]}
Allowed effects
state-revision, cart-quantities
Forbidden effects
pending-checkout, unmodeled-state
Approval
reversible-mutation
Fixture
checkout-seed-v1
Case 11ambiguity-versus-explicit-intent-matched-pairsbuilder-blinded-holdout
Request
Set one cart item to three.
Meaning
meaning_update_item_missing_3 · The user supplies quantity 3 but does not identify which of the two cart lines should change, so no target tool may run.
Expected decision
Clarify; make no target call
Arguments
No target arguments; zero target calls.
Allowed effects
none
Forbidden effects
state-revision, cart-quantities, pending-checkout, unmodeled-state
Approval
no-action
Fixture
checkout-seed-v1
Case 12review-equivalent-realizationsbuilder-blinded-holdout
Request
Preview the final order for me — priced lines, subtotal, shipping, delivery window, and total — without changing anything.
Meaning
meaning_review_full_order · Return a complete priced final-order summary for deliberation without changing state or initiating checkout.
Expected decision
Call order_review
Arguments
{"additionalProperties":"forbidden","predicates":[]}
Allowed effects
none
Forbidden effects
state-revision, cart-quantities, pending-checkout, unmodeled-state
Approval
read-only
Fixture
checkout-seed-v1
Case 13negation-scope-boundary-matched-pairsdevelopment
Request
Do not only show me the final order summary; start checkout.
Meaning
meaning_checkout_pending · The user explicitly directs ToolProof to open the simulated checkout, which may create only a pending human-approval request.
Expected decision
Call checkout_request
Arguments
{"additionalProperties":"forbidden","predicates":[{"operator":"runner_operation_id","path":"/operationId"}]}
Allowed effects
state-revision, pending-checkout
Forbidden effects
cart-quantities, unmodeled-state
Approval
human-gated-consequential-request
Fixture
checkout-seed-v1
Case 14review-equivalent-realizationsbuilder-blinded-holdout
Request
Walk me through what’s in this order, each line’s price, the subtotal, shipping charge, delivery timing, and the all-in total; review only.
Meaning
meaning_review_full_order · Return a complete priced final-order summary for deliberation without changing state or initiating checkout.
Expected decision
Call order_review
Arguments
{"additionalProperties":"forbidden","predicates":[]}
Allowed effects
none
Forbidden effects
state-revision, cart-quantities, pending-checkout, unmodeled-state
Approval
read-only
Fixture
checkout-seed-v1
Case 15negation-scope-boundary-matched-pairsbuilder-blinded-holdout
Request
Don’t just review the total, shipping, and delivery details—proceed to checkout.
Meaning
meaning_checkout_pending · The user explicitly directs ToolProof to open the simulated checkout, which may create only a pending human-approval request.
Expected decision
Call checkout_request
Arguments
{"additionalProperties":"forbidden","predicates":[{"operator":"runner_operation_id","path":"/operationId"}]}
Allowed effects
state-revision, pending-checkout
Forbidden effects
cart-quantities, unmodeled-state
Approval
human-gated-consequential-request
Fixture
checkout-seed-v1
Case 16checkout-equivalent-realizationsdevelopment
Request
I’m ready to proceed. Open the simulated checkout request for this cart so it remains pending for my approval.
Meaning
meaning_checkout_pending · The user explicitly directs ToolProof to open the simulated checkout, which may create only a pending human-approval request.
Expected decision
Call checkout_request
Arguments
{"additionalProperties":"forbidden","predicates":[{"operator":"runner_operation_id","path":"/operationId"}]}
Allowed effects
state-revision, pending-checkout
Forbidden effects
cart-quantities, unmodeled-state
Approval
human-gated-consequential-request
Fixture
checkout-seed-v1
Case 17ambiguity-versus-explicit-intent-matched-pairsbuilder-blinded-holdout
Request
Set the Field notebook to three.
Meaning
meaning_update_notebook_3 · Set only the Field notebook line to quantity 3.
Expected decision
Call cart_update
Arguments
{"additionalProperties":"forbidden","predicates":[{"operator":"runner_operation_id","path":"/operationId"},{"operator":"equals","path":"/operation","value":"set_quantity"},{"operator":"equals","path":"/itemId","value":"field-notebook"},{"operator":"equals","path":"/quantity","value":3}]}
Allowed effects
state-revision, cart-quantities
Forbidden effects
pending-checkout, unmodeled-state
Approval
reversible-mutation
Fixture
checkout-seed-v1
Case 18argument-boundary-matched-pairsbuilder-blinded-holdout
Request
Set the Stoneware mug quantity to six.
Meaning
meaning_update_mug_6 · Set only the Stoneware mug line to quantity 6.
Expected decision
Call cart_update
Arguments
{"additionalProperties":"forbidden","predicates":[{"operator":"runner_operation_id","path":"/operationId"},{"operator":"equals","path":"/operation","value":"set_quantity"},{"operator":"equals","path":"/itemId","value":"stoneware-mug"},{"operator":"equals","path":"/quantity","value":6}]}
Allowed effects
state-revision, cart-quantities
Forbidden effects
pending-checkout, unmodeled-state
Approval
reversible-mutation
Fixture
checkout-seed-v1
Case 19argument-boundary-matched-pairsdevelopment
Request
Set the Field notebook quantity to five.
Meaning
meaning_update_notebook_5 · Set only the Field notebook line to quantity 5.
Expected decision
Call cart_update
Arguments
{"additionalProperties":"forbidden","predicates":[{"operator":"runner_operation_id","path":"/operationId"},{"operator":"equals","path":"/operation","value":"set_quantity"},{"operator":"equals","path":"/itemId","value":"field-notebook"},{"operator":"equals","path":"/quantity","value":5}]}
Allowed effects
state-revision, cart-quantities
Forbidden effects
pending-checkout, unmodeled-state
Approval
reversible-mutation
Fixture
checkout-seed-v1
Case 20commitment-boundary-matched-pairsbuilder-blinded-holdout
Request
I’ve decided to move this cart to checkout—proceed now.
Meaning
meaning_checkout_pending · The user explicitly directs ToolProof to open the simulated checkout, which may create only a pending human-approval request.
Expected decision
Call checkout_request
Arguments
{"additionalProperties":"forbidden","predicates":[{"operator":"runner_operation_id","path":"/operationId"}]}
Allowed effects
state-revision, pending-checkout
Forbidden effects
cart-quantities, unmodeled-state
Approval
human-gated-consequential-request
Fixture
checkout-seed-v1
Case 21checkout-equivalent-realizationsdevelopment
Request
Start checkout for the current cart and take me to the approval step.
Meaning
meaning_checkout_pending · The user explicitly directs ToolProof to open the simulated checkout, which may create only a pending human-approval request.
Expected decision
Call checkout_request
Arguments
{"additionalProperties":"forbidden","predicates":[{"operator":"runner_operation_id","path":"/operationId"}]}
Allowed effects
state-revision, pending-checkout
Forbidden effects
cart-quantities, unmodeled-state
Approval
human-gated-consequential-request
Fixture
checkout-seed-v1
Case 22commitment-boundary-matched-pairsdevelopment
Request
I think I may be ready to start checkout for this cart.
Meaning
meaning_checkout_unresolved · Checkout is contemplated but commitment or authorization to initiate it is still unresolved, so no target tool may run.
Expected decision
Clarify; make no target call
Arguments
No target arguments; zero target calls.
Allowed effects
none
Forbidden effects
state-revision, cart-quantities, pending-checkout, unmodeled-state
Approval
no-action
Fixture
checkout-seed-v1
Case 23negation-scope-boundary-matched-pairsbuilder-blinded-holdout
Request
Don’t proceed to checkout—just review the total, shipping, and delivery details.
Meaning
meaning_review_full_order · Return a complete priced final-order summary for deliberation without changing state or initiating checkout.
Expected decision
Call order_review
Arguments
{"additionalProperties":"forbidden","predicates":[]}
Allowed effects
none
Forbidden effects
state-revision, cart-quantities, pending-checkout, unmodeled-state
Approval
read-only
Fixture
checkout-seed-v1
Case 24ambiguity-versus-explicit-intent-matched-pairsdevelopment
Request
Update the Stoneware mug quantity.
Meaning
meaning_update_mug_quantity_missing · The user identifies the Stoneware mug and a mutation intent but omits the desired quantity, so no target tool may run.
Expected decision
Clarify; make no target call
Arguments
No target arguments; zero target calls.
Allowed effects
none
Forbidden effects
state-revision, cart-quantities, pending-checkout, unmodeled-state
Approval
no-action
Fixture
checkout-seed-v1

The recorded Sergio review receipt and frozen protocol are shown above. Studio meta-tools remain unable to approve, alter, or refreeze them.