decision ledger

{
  "schema_version": "actionist.decision-ledger.v1",
  "part": "P13",
  "run_id": "2026-08-27-sprint-1-fable",
  "recorded": "2026-08-27",
  "owner": "ACTIONIST-S1-L5-RUNTIME",
  "boundary": {
    "research_only": true,
    "execution_status": "UNEXECUTED",
    "admission_status": "NOT_ADMITTED",
    "admitted_blocks": 0,
    "implementation_authorized": false,
    "promotion": "unpromoted"
  },
  "decisions": [
    {
      "id": "D-P13-01",
      "statement": "Bounded editing is limited by the least-bounded channel, not by the edit verb set",
      "state": "accepted",
      "support": "30-surface census: all seven verbs near-universal; three boundary postures; every AI builder shows unbounded fallthrough",
      "falsifier": "A system with an ungated chat channel nevertheless preserves compositional properties under audit",
      "evidence_class": "observed_behavior"
    },
    {
      "id": "D-P13-02",
      "statement": "Store intent, not mutation; upgrade safety is a property of what was stored",
      "state": "accepted",
      "support": "No surveyed surface durably records why; coordinate-bound edits cannot survive node movement",
      "falsifier": "Stored intents fail to re-resolve on a real minor-version upgrade",
      "evidence_class": "observed_behavior"
    },
    {
      "id": "D-P13-03",
      "statement": "Out-of-bounds edits must be unrepresentable, not merely blocked",
      "state": "accepted",
      "support": "Storyblok whitelists and Power Apps managed layers are the only architecturally sound posture observed",
      "falsifier": "A donor whose editable surface cannot be typed without rewriting it",
      "evidence_class": "observed_behavior"
    },
    {
      "id": "D-P13-04",
      "statement": "Layer separation is necessary but not sufficient for upgrade safety",
      "state": "accepted",
      "support": "Verified: Shopify OS2.0 settings layer reportedly wiped on theme update, Mar 2026, two independent merchant reports plus staff acknowledgement of receipt",
      "falsifier": "A per-upgrade snapshot/diff gate proves unnecessary across a representative donor set",
      "evidence_class": "observed_behavior"
    },
    {
      "id": "D-P13-05",
      "statement": "Adopt `deferred` as a fourth edit outcome",
      "state": "accepted",
      "support": "Local siso-ui-base gate reports DNA-overridden values as deferred without failing the run",
      "falsifier": "Clients cannot distinguish deferred from denied in practice",
      "evidence_class": "first_party_docs"
    },
    {
      "id": "D-P13-06",
      "statement": "Open Design is not a live foundation",
      "state": "accepted",
      "support": "opendesigndev org last activity 3 Sep 2024; nothing 2025-2026; open-design.ai is an unrelated agent wrapper; ODA is unrelated CAD/BIM",
      "falsifier": "Evidence of active 2026 maintenance on the Open Design Engine or Octopus spec",
      "evidence_class": "observed_behavior"
    },
    {
      "id": "D-P13-07",
      "statement": "Data-binding is the highest-risk verb and must be scope-gated",
      "state": "accepted",
      "support": "Only verb that alters what data the app touches; market UIs weight it like a style change",
      "falsifier": "A binding model that provably cannot widen scope regardless of gating",
      "evidence_class": "inference"
    },
    {
      "id": "D-P13-08",
      "statement": "Commercial denominator is 30, not ~100 — depth contract unmet",
      "state": "recorded_gap",
      "support": "Account-wide quota terminated the sweep; no padding performed",
      "falsifier": "A completed continuation wave closes it",
      "evidence_class": "observed_behavior"
    }
  ]
}