PROMPT 1 — P29.1a — GOVERNED EFFECT IDENTITY / RETRY / PROTE...

Prompt

PROMPT 1 — P29.1a — GOVERNED EFFECT IDENTITY / RETRY / PROTECTIVE ACTION / CURRENTNESS Mode: EXACT-PAYLOAD / SEARCH-OFF / P-ONLY / DESIGN-AUDIT INPUT BOUNDARY Use only this exact payload. Do not use Web/Search, prior Prompt 1 history, M/Z, sibling outputs, provider identity, hidden platform state or implementation traces. Mechanisms below are design text, not runtime proof. Return only the required final sections. UNKNOWN != ABSENT; PARTIAL != TOTAL; DESIGN-SOUND != IMPLEMENTED. EVALUATION RULES - Exact counterexample and decision consequence outrank vote count, prestige, verbosity or confidence. - Use the strongest conservative reading: reject a claimed defect if an explicit clause closes it; retain a defect only if a compliant bypass, ambiguity or decision-incomplete state survives. - Prefer REUSE -> MERGE -> EXTEND -> NEW. New control IDs require a genuinely distinct invariant/failure mechanism. - Same-model/same-data/same-state/repeat agreement is not independent confirmation. Freshness != independence. - Labels, scores, confidence, convergence states and audit outputs never create authority. - Positive identity/currentness/independence/isolation/parity/containment claims require positive evidence. - Protective/non-weakening action may remain fast; weakening or authority expansion is governed. - A valid NO-CHANGE is preferable to mechanism growth without decision value. FIXED GLOBAL INVARIANTS 1. Safety, truth and epistemic integrity outrank utility, speed and optimization. 2. No self-attestation, relabel, retry, evaluator swap, majority or confidence alone upgrades a restricted state. 3. Blocking is dependency-scoped; unrelated safe/read-only/protective work may continue. 4. Material transitions are typed, lineage-bound, currentness-aware and rollback/reopen-aware where feasible. 5. Candidate-generation evidence cannot validate the candidate that generated it. 6. No audit result itself authorizes PROMOTE/INTEGRATE/RELEASE. AUDIT TARGET — SELF-CONTAINED DOSSIER A. GOVERNED PROTECTION-REDUCING DETERMINATION A1. Any determination that can reduce protection — initial risk/tier, materiality, N/A, exception, scope/completeness reduction, blocker limitation/closure, restoration or renewal — uses a typed record: object/scope; prior/proposed state; evidence/provenance; action class; beneficiary/conflict; authorized role; required independence/SOD; expiry/recheck; lineage; allowed decision-use; rollback/reopen consequence. Missing required material field => UNKNOWN => dependent material action HOLD/BLOCK. A2. A benefiting path cannot reduce its own protection by relabeling the action LOW, editorial, benign, N/A or non-weakening. UNKNOWN/ambiguous class defaults to the higher plausible protection until positively reduced by A1. A3. Pure HOLD/QUARANTINE/protection-increasing action may use a fast path with lineage. Any composite containing a weakening component routes through A1. A4. Standing/material exceptions are narrow, currentness-bound and non-resetting. Renewal is a new governed determination with fresh evidence and cumulative exposure; novelty of rationale is not required merely for novelty. B. INTENT / EFFECT IDENTITY B1. INTENT-OCCURRENCE-ID is minted when a governed intent is created. EFFECT-IDENTITY-KEY binds that intent occurrence + real-world target + operation semantics + governed scope. Attempt/request IDs are mutable and cannot detach the same intended effect from unresolved state. B2. Distinct genuine intents may have byte-identical payloads. If distinctness is ambiguous, inherit same-effect duplicate-risk until authoritative reconciliation. B3. COMMIT-INDETERMINATE/duplicate-risk follows the intended effect across attempts. Actor-only claims cannot upgrade outcome or idempotency. B4. Every retry/commit atomically revalidates current authority/policy, tier/materiality, exception validity, blockers, support/currentness, identity, fencing/idempotency and containment. C. OUTCOME-SENSITIVE RETRY C1. EXECUTED forbids producing the same effect again. C2. NOT-EXECUTED permits a new attempt only when terminal nonexecution and exclusion of stale attempts are positively established, followed by full current-prerequisite revalidation. C3. PARTIAL / CONFLICT / UNKNOWN / STALE holds the affected effect unless a current authoritative effect-side key-bound deduplication mechanism prevents a new effect; deduplication prevents duplication but does not supply authority/current evidence/containment. C4. Cancellation or fencing after possible commit does not prove nonexecution. Reissuing a request/attempt ID does not prove a new intent. D. CURRENTNESS / DEPENDENCY D1. Material support invalidation or expiry reopens/restricts actual dependents. Unrelated safe work remains available. D2. Cumulative exposure includes sibling exceptions and successor baselines, not only renewals of one record. D3. Unknown material oversight/currentness predicates restrict dependent promotion without being relabeled as empirically DEGRADED. STRESS TESTS A. Benefiting path labels a high-consequence action LOW before any strict gate. B. Material exception omits expiry and is used months later. C. Same intended payment times out; actor mints a new attempt ID and retries. D. Actor calls an unresolved same-target operation a “new intent” to shed COMMIT-INDETERMINATE. E. Two positively distinct purchases have byte-identical payloads and both must remain possible. F. Actor-only “idempotent” assertion is used while authoritative effect reconciliation is unavailable. G. Attempt 1 was authorized; before retry the exception expired and evidence was invalidated. H. Two blockers close but a third load-bearing blocker remains. I. Protective QUARANTINE is delayed because a weakening approver is unavailable. J. QUARANTINE plus tier downgrade is sent through the protective fast path. K. Same rationale is used for a renewal with fresh current evidence. L. A decision remains active after its load-bearing support crosses recheck trigger. M. Commit completes just before fencing; reconciliation says EXECUTED; actor asks for same effect again. N. Terminal NOT-EXECUTED is known but one stale attempt can still execute. O. Outcome is UNKNOWN, but a qualified same-key effect-side deduplication service and all current prerequisites are positive. P. Two distinct intents share payload and target type but have independently established occurrences. Q. Sibling exceptions jointly exceed the qualified cumulative exposure while each local record is current. R. A protective-only HOLD is applied immediately with trace lineage while full adjudication follows. REQUIRED OUTPUT — EXACT ORDER 1. EXECUTIVE VERDICT — 3–6 sentences; exactly one dominant NEXT ACTION. 2. AUDIT TRACE / REJECTED CANDIDATES / UNCERTAINTIES. 3. SECTION RATINGS — each major block DESIGN-SOUND / DESIGN-DEFECT / UNVERIFIED / N/A with one-sentence reason. 4. STRESS-TEST MATRIX — every letter A–R -> BLOCKED / BYPASS / UNRESOLVED / ALLOWED-BENIGN; cite mechanism + exact clause. BLOCKED is textual/design prevention, not observed runtime interception. 5. TOP DOMINANT FINDINGS — max five; for each retained material defect give claim, exact clause, impact, root cause, reproducible path, minimal repair, new risk/complexity, falsifiable validation and disposition. 6. REDUNDANCY / MERGE CANDIDATES + concise residual register. 7. MISSING-CONTROL CANDIDATES — only if not encoded or directly mergeable. 8. RECOMMENDATION SET — max five; MERGE / EXTEND / NEW / REJECT / DEFER with falsifiable validation. 9. FINAL SCOPE STATEMENT — exact-P design audit only; implementation effectiveness UNVERIFIED; no release authority.