General Proposal · Proposed / Non-Authoritative
Direction Level · Single-file HTML HDO variance

T104-PH001-AC005-CHECK000 commissioning decision proposal

Five pending decisions on internal Detail depth, Decision-card boundaries, concrete content roles, proposal-file separation, and the first formal task carrier route for general_proposal.

Checkpoint
T104-PH001-AC005-CHECK000
Review cycle
Post-escalation rebuild
Version
v0.5.0
Date
2026-07-12
Status
Draft · pending CHECK000 disposition
Decision owner
HITL (HDO)
Producer
AFK (LLM_Consultant)
Authority boundary. This file claims no disposition or execution authority. Selections and notes are non-persisting routed input that the HDO may copy out for Consultant scribing. Disposition happens only through the P-STD-008-CLAUSE-005A four-leg bundle on commissioning PR #493; this file, its recommendations, and its payload do not substitute for that record.
Architecture decisions · AD0 / 4 dispositioned
Governance decisions · GD0 / 1 dispositioned
1
Orientation
What is being decided

The HDO is reviewing five pending decisions in two independent lanes. Decisions 1–4 form the Architecture lane for Detail-depth and proposal-card mechanics. Decision 5 forms the Governance lane for the TK001 carrier route. Nothing is pre-approved, and no GD state implies an AD state or vice versa.

Architecture lane · 0 / 4
Decisions 1–4
Internal Detail semantics · Decision-card boundary · three content roles · one-file/two-file separation
Governance lane · 0 / 1
Decision 5
Amend #326 in place · fresh same-UID carrier · fresh child UID
2
Grounded constraints
What the SSOT establishes

Primary grounding: artifacts/tasks/T104/workspace/PH001/AC005/analysis/analysis_T104-PH001-AC005-CHECK000_detail-depth-decision-grounding.md, especially §III.B. These are established constraints and recorded lineage, not dispositions.

Standards, PRD, and SPS

  • T104-STD-005-CLAUSE-003 requires citation rather than absorption of other artifact families; CLAUSE-005 keeps detail reserved-inactive; CLAUSE-008, CLAUSE-009, and CLAUSE-011 preserve proposal authority and activation boundaries.
  • The #98 standard↔PRD contract activates only direction and architecture and requires recommendation markers plus pending decision-level controls.
  • sps_T104C §xi defines Detail as supporting evidence, tradeoff, constraint, or implementation cue under an Option; Decision is a CEO-reviewable question inside a Direction.

Direction lineage · compared landscape

  • COMM c5 (#314), originating in the 2026-06-17 CBI (#308), requested richer code/schema/diagram depth at the internal Details layer and a boundary around process material.
  • Parent PLAN #325 and existing issue #326 record the later precommission-first, thin-formal-slice correction.
  • Compared candidates: precommission then execute and thin exact formal slices remain carried; existing diagnostic TK001 as-is is not carried; reserved proposal_level=detail activation remains rejected for this pass.
  • AC008 demonstrates bounded internal Detail exhibits and independent AD/GD lanes, but that dogfood is qualitative, non-authoritative evidence and does not decide AC005.
3
Prompt-specific judgment
What remains Consultant synthesis

Recommendations 1A–4A

1A answers c5 without activating reserved altitude vocabulary. 2A prevents the catch-all failure monitored by T104C-RISK-003 while preserving citations. 3A makes explanation, illustration, and candidate downstream text visibly distinct; the proposed carry-forward role is new synthesis and has the largest TK002 wording impact. 4A applies T104C-CON-006 as a subject-identity test: one subject per file, split only when owner, authority path, timing, or disposition genuinely diverges.

Decision 5 · honest tension

5A remains recommended because the pre-freeze commissioning package can defensibly define TK001’s refined contract with the lowest churn and a CLAUSE-008D changelog trail. 5D preserves the same UID with a fresh issue carrier. 5E is the strictest append-only reading: a new deliverable target requires a fresh child UID and P-STD-005-CLAUSE-007B forbids reuse of retired TK001. The SSOT supports more than one conforming answer; only the HDO adjudicates the weight.

4
Carrier-route comparison
TK001 route mechanics · Decision 5
Criterion5A · amend #326 in place5D · fresh same-UID carrier5E · fresh child UID
Commissioned-contract readingRefinement within the pre-freeze targetRe-scoped continuation with carrier replacementNew deliverable target
P-STD-008-CLAUSE-008CIn-place refinement is defensibleRetire #326; preserve TK001 identityRetire TK001; mint a new child UID
P-STD-008-CLAUSE-008DBody edit plus what/why changelog commentPointer closure plus fresh-carrier lineageRetirement pointer plus new-UID lineage
P-STD-005-CLAUSE-007BNo UID retirement or reuseSame task UID continues on a new issue carrierRetired TK001 is never reused
Cost / integrity postureLowest churn; strongest continuityMore issue churn; same task identityStrictest append-only posture; new task identity
Consultant posture: 5A is recommended, but this is not a numeric winner or a pre-disposition. GAP-003 remains a genuine interpretation tension. The HDO may select 5D or 5E without contradicting the established record, provided the corresponding retirement, lineage, freeze, and mint mechanics are followed.
5
Architecture decision lane · AD
Detail-depth and proposal-card dispositions

Four independent, pending architecture decisions. The green recommendation treatment and ✅ identify the Consultant recommendation only; no disposition state is pre-filled.

Filter

Architecture lane · AD

Detail-depth mechanics · 4 pending decisions

D1Detail-depth semanticsPending
Should Detail-depth enrich internal Detail content, or activate reserved proposal_level=detail?

Recommendation

✅ 1A · Richer internal Detail content. Keep active proposal levels at direction and architecture; allow richer technical support inside internal Details.

Options and consequence deltas

1B · Activate proposal_level=detail. Treat c5 as a request to make detail an authorable proposal altitude. If chosen instead, ordinary carry-forward stops until the proper PRD/standard authority activates the reserved value.

1C · Leave depth ad hoc. Avoid mechanics changes and rely on author judgment. If chosen instead, COMM c5 remains unresolved and internal-depth consistency stays unenforceable.

Reasoning

T104-STD-005-CLAUSE-005 reserves detail as inactive; CLAUSE-011 routes activation to #98. sps_T104C §xi already defines Detail as an internal content unit.

Downstream impact · High

If 1A is accepted, #98 and proposal guidance/template wording can define richer internal Detail content; proposal_level=detail remains inactive, and any P-STD-005 impact routes later rather than being edited here.

Established vocabulary · Consultant 1A synthesis
Non-authoritative illustrative Detail · bounded internal depth

Illustration only. A Detail may name edge conditions, affected fields, and failure behavior beneath an Option. This is not implementation text and does not create a proposal altitude.

detail: { depth: bounded, supports_option: true, authority: none }
Evidence boundary
AC008’s labeled collapsible exhibits show that bounded Detail is workable, but that sibling dogfood is qualitative and non-authoritative for AC005.
D2Decision-card boundaryPending
What belongs in a general_proposal Decision card when governance or process facts affect the choice?

Recommendation

✅ 2A · Technical/architecture cards; cite process context. Decision cards carry technical or architecture choices. Governance/process context is cited or handled in the owning plan, notes, gate, PMB, or issue surface.

Options and consequence deltas

2B · Mixed technical and process ledgers. Cards may directly host process routing, ownership, PMB, or approval-flow choices. If chosen instead, general_proposal risks becoming a catch-all control record and duplicating owning surfaces.

2C · Remove process context entirely. Cards ignore governance/process context even when it explains the recommendation. If chosen instead, material constraints and the evidence trail are lost.

Reasoning

T104-STD-005-CLAUSE-003 and T104C-QG-004 require cite-don’t-absorb discipline; T104C-RISK-003 monitors the catch-all failure mode.

Downstream impact · High

If 2A is accepted, guidance separates architecture decision groups from governance/process groups while gate/checkpoint and PMB decisions remain in their owning surfaces.

Boundary doctrine · Consultant 2A synthesis
Evidence boundary
AC008 shows independent AD/GD presentation can preserve context without turning architecture cards into process ledgers; it does not pre-approve this choice.
D3Detail, Example, and Proposed carry-forward rolesPending
How should concrete material remain useful to the HDO without being mistaken for approved execution authority?

Recommendation

✅ 3A · Three distinct labeled roles. Detail explains the option; Example illustrates; Proposed carry-forward detail is candidate downstream text/spec that must be explicitly promoted later.

Options and consequence deltas

3B · Treat all concrete material as Detail. Simpler vocabulary, but examples and candidate specifications can appear approved. If chosen instead, readers may mistake illustration for implementation authority.

3C · Ban concrete material. Prevents false authority but weakens HDO review of architecture and makes design tradeoffs less testable.

Reasoning

The SPS defines Detail but does not name Proposed carry-forward detail. The added role makes explicit that candidate text acquires authority only after promotion into its owning PRD/spec/plan/task surface.

Downstream impact · Medium

TK002 guidance/template wording gains explicit labels. Historical proposals remain point-in-time records; accepted candidate text still must be copied or transformed into an authorized downstream artifact.

Largest downstream wording impact · Consultant synthesis
Non-authoritative illustrative Detail · three roles

Illustration only. These samples demonstrate labeling, not approved implementation material.

Detail
Explains
Evidence, tradeoff, constraint, or implementation cue under an Option.
Example / carry-forward
Illustrates / proposes
An Example never prescribes. Proposed carry-forward text remains a candidate until explicitly promoted.
Evidence boundary
T104C-CON-007 and T104-STD-005-CLAUSE-009 keep all three roles non-authoritative until the governed disposition and any later promotion.
D4One-file versus two-file separationPending
When should architecture and governance/process review needs share one proposal file, and when should they split?

Recommendation

✅ 4A · Split only when review needs materially differ. Keep one proposal when process context is only evidence; split into two proposal subjects when architecture and governance/process decisions require different dispositions.

Options and consequence deltas

4B · Always one file. Preserve one review packet even when decision groups differ. If chosen instead, process-heavy review can obscure the architecture decision the HDO needs to make.

4C · Always split files by category. Force separate files even for minor process context. If chosen instead, navigation and synchronization costs rise without a distinct subject or review contract.

Reasoning

T104C-CON-006 makes separation a subject-identity rule. One file is valid when the subject belongs together and lanes remain independent; split when owner, authority path, review timing, or disposition genuinely diverges.

Downstream impact · Medium

Guidance gains a separation test and examples; SP3 may dogfood the split under draft/non-authoritative labeling, while parent plans and session notes retain control sequencing.

CON-006 constraint · Consultant 4A synthesis
Evidence boundary
AC008 proves the one-file branch is workable; it is sibling evidence, not authority and not a pre-made AC005 disposition.
6
Governance decision lane · GD
TK001 carrier-route disposition

One pending governance decision. The three operable routes preserve the genuine GAP-003 tension; 5A is recommended, while 5D and 5E remain conforming alternatives with different lineage costs.

Governance lane · GD

Carrier and mint mechanics · 1 pending decision

D5TK001 lineage, issue #326 treatment, and child-task decompositionPending
How should the pre-freeze issue #326 lineage become the first thin execution slice, and when may the selected child-task set mint?

Recommendation

✅ 5A · Amend issue #326 in place under CLAUSE-008D. Keep #326 as the carrier for T104-PH001-AC005-TK001; after conforming approving CHECK000 proof, amend its body to the frozen Draft 1 PRD-authority contract with the P-STD-008-CLAUSE-008C/008D changelog-comment trail.

Operable alternatives and consequence deltas

5D · Retire issue #326 and mint a fresh carrier under CLAUSE-008C. At commissioning, close #326 as retired-unexecuted with a pointer comment, then mint a fresh carrier issue for the same T104-PH001-AC005-TK001 UID carrying Draft 1. This preserves the re-scoped-continuation reading but adds issue churn; it does not create a new task identity.

5E · Retire TK001 and mint a fresh child task UID under CLAUSE-008C. Close #326 as retired-unexecuted with a pointer comment, retire T104-PH001-AC005-TK001 without reuse, and mint a fresh child UID (candidate T104-PH001-AC005-TK007) carrying Draft 1. This is the strictest new-deliverable-target reading; P-STD-005-CLAUSE-007B forbids reuse.

Reasoning

5A is lowest-churn and defensible because commissioning defines the pre-freeze task contract. 5E is the strictest append-only interpretation. Only the HDO can weight continuity against the new-target reading.

Downstream impact · High

Under 5A/5D the freeze covers TK001–TK005. Under 5E it covers TK002–TK005 plus the fresh UID. TK000 and TK006 remain retired pre-freeze. No issue is created, amended, retired, or executed by this HTML.

GAP-003 · genuine HDO adjudication
Non-authoritative Detail · route consequence map

Illustration only. Mint timing becomes actionable only after the complete approving CHECK000 record.

5A#326 continues as TK001; body + changelog comment.
5D#326 retires; fresh issue carries same TK001 UID.
5ETK001 retires; fresh child UID carries Draft 1.
Evidence boundary
#325 records the re-scope intent; #326 remains diagnostic. CLAUSE-008C/008D distinguish refinement from new-target minting, and P-STD-005-CLAUSE-007B prevents retired UID reuse.
7
Never hidden
Open residue

Only unresolved items remain here; resolved review history stays in provenance.

D5 weight

The HDO must adjudicate how much weight to give continuity/low churn (5A), fresh-carrier lineage (5D), and the strict new-deliverable/append-only reading (5E).

Ledger

After an approving merge, the acceptance-ledger comment on parent PLAN issue #325 must bind the outcome to the merge hash, commissioning PR #493, and this decision surface.

Freeze

The exact register freezes only through the approving commissioning record: TK001–TK005 under 5A/5D, or TK002–TK005 plus the newly minted child UID under 5E.

Mint timing

Post-merge issue amendment, retirement, carrier creation, and follow-on mint timing remain governed mechanics; none is executed or authorized by this proposal.

8
Routed input only
HDO controls and scribe payload

All five controls ship pending. Selections, alternative selections, and notes remain in memory only. “Change” requires an alternative selection. Readiness stays unset until all five decisions have a disposition; all Accept yields ready_to_carry_forward, while any Change/Other yields recycle.

Architecture lane

0 Accept · 0 Change · 0 Other · 4 pending

Governance lane

0 Accept · 0 Change · 0 Other · 1 pending

Leg 1

The merge event is the anchoring auditable platform event.

Leg 2

Commissioning PR #493 carries the CHECK000 control-node record, all five dispositions, evidence links, and the recorded HDO act.

Leg 3

The parent PLAN issue #325 receives the merge-hash-bound acceptance-ledger comment.

Leg 4

The decision-record linkage propagates the five dispositions into the drift-audit §V.A HDO register; any authorized session-note linkage strengthens but does not replace the bundle.

P-STD-008-CLAUSE-005A requires all four linked legs. Copying this payload is user-initiated egress for governed scribing only; it is not any leg of the disposition record by itself.

9
Traceability
Provenance and authorization boundary

Lineage: retired Markdown cycles 1–3 → WDL baseline-drift audit → adversarial rounds 0/1 → v0.4.0 post-escalation rework → this v0.5.0 exemplar-skeleton rebuild. The review campaign and resolved history are provenance, not open residue.

MD cycles 1–3
Precommissioning
Decision substance, compared landscape, and impact deltas stabilized before retirement in place.
WDL drift audit
Reconciliation
Control mechanics, GAP register, and the §V.A HDO decision register were reconciled.
Rounds 0 / 1
Adversarial recycle
Two review rounds surfaced instrument and carrier-route defects; HDO authorized post-ceiling remediation.
v0.4.0 → v0.5.0
Family rebuild
Decision substance is re-instrumented on the stabilized AC008 exemplar skeleton for HDO visual review.
artifacts/tasks/T104/workspace/PH001/AC005/plan_T104-PH001-AC005.md
artifacts/tasks/T104/workspace/PH001/AC005/analysis/analysis_T104-PH001-AC005-CHECK000_detail-depth-decision-grounding.md
artifacts/tasks/T104/workspace/PH001/AC005/analysis/analysis_T104-PH001-AC005_wdl-baseline-drift-audit.md §V.A
https://github.com/NMAQ1607/prompt/pull/493

Governance footnotes

  • T104-STD-005-CLAUSE-003: cite, do not absorb, other artifact families.
  • T104-STD-005-CLAUSE-009: proposal content remains non-authoritative until the governed disposition.
  • DRC remains decision-level; readiness is advisory and never authorizes execution.

Authorization boundary

This proposal amends no standard, PRD, guideline, template, issue, register, or board state. It does not activate proposal_level=detail. The four-leg proof bundle on PR #493 is the only CHECK000 disposition path described here.