Context
Sibling leaf of the target-architecture composition ADR (graduated from Discussion #13846), split out per @neo-gpt's graduation constraint: the identity-substrate alignment must be governed, never a stealth rider inside an architecture PR. Live latest-open sweep at 2026-07-03T00:55:29Z: no equivalent ticket.
The Problem
AGENTS.md — the always-loaded session substrate — still asserts "four co-load-bearing pillars" in TWO anchors: the top §core_values bridge text and §neo_identity_anchor. That is STALE against the operator's own decision (ADR-0018 OD-3, 2026-05-30): Two Hemispheres is the canonical top-level scaffold (Body /src/ ↔ Brain /ai/, joined by the Neural Link); Four Pillars is demoted to Brain-internal elaboration. Every session (including this one — verified live) boots carrying an identity frame the identity authority already overruled. ADR-0018 §2 itself notes AGENTS.md "had self-contradictory pillar wording."
The Fix
A narrow, governed edit repairing BOTH anchors to the two-hemisphere-canonical framing (four-pillar retained only as the Brain-internal elaboration it now is), executed under:
- ADR-0018 OD-3 (the deciding authority — this is a staleness fix, not a re-decision),
neo-identity-update (the identity-surface change protocol; AGENTS.md is one of the ~30 identity-encoding surfaces),turn-memory-pre-flight (AGENTS.md is loaded instruction substrate: the PR carries the slot-rationale section — disposition rewrite on both anchors, net byte movement stated).
Acceptance Criteria
Out of Scope
- The composition ADR itself (the sibling authoring leaf).
- Re-deciding the scaffold (ADR-0018 OD-3 already did).
Related
Related: Discussion #13846 (constraint origin), ADR 0018 (authority), the composition-ADR authoring leaf (sibling)
Origin Session ID: 8cf234b7-e698-47ca-99e2-bf865196b6aa
Retrieval Hint: AGENTS.md four pillar two hemisphere alignment identity anchor ADR-0018 OD-3 governed
Context
Sibling leaf of the target-architecture composition ADR (graduated from Discussion #13846), split out per @neo-gpt's graduation constraint: the identity-substrate alignment must be governed, never a stealth rider inside an architecture PR. Live latest-open sweep at 2026-07-03T00:55:29Z: no equivalent ticket.
The Problem
AGENTS.md— the always-loaded session substrate — still asserts "four co-load-bearing pillars" in TWO anchors: the top§core_valuesbridge text and§neo_identity_anchor. That is STALE against the operator's own decision (ADR-0018 OD-3, 2026-05-30): Two Hemispheres is the canonical top-level scaffold (Body/src/↔ Brain/ai/, joined by the Neural Link); Four Pillars is demoted to Brain-internal elaboration. Every session (including this one — verified live) boots carrying an identity frame the identity authority already overruled. ADR-0018 §2 itself notes AGENTS.md "had self-contradictory pillar wording."The Fix
A narrow, governed edit repairing BOTH anchors to the two-hemisphere-canonical framing (four-pillar retained only as the Brain-internal elaboration it now is), executed under:
neo-identity-update(the identity-surface change protocol; AGENTS.md is one of the ~30 identity-encoding surfaces),turn-memory-pre-flight(AGENTS.md is loaded instruction substrate: the PR carries the slot-rationale section — dispositionrewriteon both anchors, net byte movement stated).Acceptance Criteria
neo-identity-updatesurface checklist consulted; any OTHER surface still carrying four-pillar-as-top-level is listed (fix or follow-up per scale)Resolvesthis leafOut of Scope
Related
Related: Discussion #13846 (constraint origin), ADR 0018 (authority), the composition-ADR authoring leaf (sibling)
Origin Session ID: 8cf234b7-e698-47ca-99e2-bf865196b6aa
Retrieval Hint:
AGENTS.md four pillar two hemisphere alignment identity anchor ADR-0018 OD-3 governed