Studio executive + coverage frontier product (BL-62)

Versioned executive verb catalogue with BL-52 route pointers and a coverage-frontier score (honesty × answer-rate). Never invent-green full coverage while refuse rates are hidden.

Module: scpn_quantum_control.studio_executive_product

This page documents a bounded executive catalogue and an honesty-aware coverage metric. It does not dispatch a verb, grant approval, redesign Studio, or convert a coverage score into evidence of route completeness.

Contract discovery

Function	Contract
list_executive_verb_ids()	Returns all nine stable verbs in catalogue order.
get_executive_verb(verb_id)	Resolves one exact row; blank and unknown ids raise ValueError.
iter_executive_verbs(...)	Returns all rows or filters by support posture.
map_studio_executive_public_surfaces()	Identifies the product, ambient verb spine, and coverage-frontier owner.

When the optional Studio platform is importable, the catalogue reads ambient verb contracts. Otherwise it uses the explicit product-local fallback with the same nine bounded verbs. Discovery itself executes no verb or hardware route.

Public value objects

• ExecutiveVerbRow records route and BL-53 pointers, support posture, approval/live-hardware flags, backends, and the shared claim boundary.
• PathEligibilityDecision records an allowed/refused outcome, reason, ordered blockers, and non-promotional boundary.
• MaterialisedCoverageFrontierProbe records ledger counts, answer/honesty rates, frontier score, off-frontier state, and the mandatory false invent-green flag.

All records are immutable slot-backed dataclasses with validated construction and JSON-ready to_dict() mappings.

Rules

Rule	Behaviour
Product schema	studio_executive_product.v1
Default verb	differentiate
Unknown verb	Fail closed
Unsupported route invent-green	Refused (BL-52/53)
Full coverage invent-green	Refused
Live execute without approval	Refused

Executive path decisions

decide_executive_path() validates the exact verb, then refuses unsupported route claims, invented full coverage, and approval-gated verbs without explicit approval. Multiple blockers are de-duplicated in first-seen order.

An allowed decision says only that the product policy admits that route. It does not execute the route, authorize a provider, spend shots, or certify an output.

Claim boundary:

Studio executive + coverage frontier product surface only; catalogues executive verbs with BL-52 route pointers and materialises honesty×answer-rate coverage-frontier probes; invent_green_full_coverage=false when boundary abstentions exist; refuses invent-green unsupported routes and hidden refuse rates; does not claim full BL-55 kit export or Studio UI redesign (S62.4 residual)

Public API

from scpn_quantum_control.studio_executive_product import (
    assert_studio_executive_product_integrity,
    build_studio_executive_product_registry,
    decide_executive_path,
    list_executive_verb_ids,
    materialise_demo_coverage_frontier_probe,
)

assert "differentiate" in list_executive_verb_ids()
reg = assert_studio_executive_product_integrity(
    build_studio_executive_product_registry()
)
probe = materialise_demo_coverage_frontier_probe()
assert probe.invent_green_full_coverage is False
assert abs(probe.frontier_score - 0.24) < 1e-9  # 0.8 * 0.3

assert decide_executive_path("execute").allowed is False  # needs approval
assert decide_executive_path("execute", approval_present=True).allowed is True

Verbs (S62.0)

compile · simulate · analyse · validate · benchmark · replay · differentiate · mitigate · execute (gated live hardware)

Every verb carries a BL-52 governed-route pointer and BL-53 unsuitable-scenario pointer. Only execute may advertise live hardware, and it must require approval. The remaining verbs are local research routes in this product.

Coverage frontier (S62.3)

Demo partition: total=10, answered=3, honest abstentions=5, improvable=2 → answer_rate=0.3, honesty_rate=0.8, frontier_score=0.24, off_frontier=True.

compute_coverage_frontier_score() requires a positive total, non-negative improvable count, and a partition that does not exceed the total. It computes:

• answer_rate = answered_confident / total_claims;
• honesty_rate = (answered_confident + honest_abstentions) / total_claims;
• frontier_score = honesty_rate * answer_rate.

The result always sets invent_green_full_coverage=False. Honest abstention is bookkeeping, not an answered claim, and therefore cannot inflate answer rate.

Registry integrity

build_studio_executive_product_registry() emits schema studio_executive_product.v1, the full verb catalogue, public surface map, default id, counts, policy note, and shared claim boundary.

Always validate transported or stored payloads through assert_studio_executive_product_integrity(). It rejects missing, empty, non-list, non-mapping, blank, duplicate, missing, or extra rows; missing route pointers or backend lists; live hardware on any verb except execute; loss of the differentiate or execute sentinels; count drift; and any relaxed invent-green policy.

Failure handling and operational non-effects

Treat ValueError as a caller-contract, route-policy, score-partition, or transported registry failure. Treat RuntimeError from catalogue construction as repository or optional-platform contract corruption.

This product performs no network access, credential lookup, provider or QPU discovery, verb execution, approval mutation, hardware submission, spend, benchmark, evidence export, BL-55 kit generation, Studio UI mutation, result promotion, or publication.

Bounded product status

Shipped: S62.0 verb→route catalogue · S62.1 inventory · S62.2 fail-closed verb paths · S62.3 coverage frontier metric + tests · S62.5 product docs.

Open residual: S62.4 evidence bundle export → BL-55 kit.

Authored by Anulum Fortis & Arcane Sapience (protoscience@anulum.li)