spec(reasoning): wire the Assay into ReasoningReceipt (run-level epistemic posture)

CHANGELOG.md

@@ -10,6 +10,7 @@ The format follows [Keep a Changelog](https://keepachangelog.com/en/1.0.0/). Thi
 
 ### Added
 - Epistemic Assay contracts: `ReasoningAssay` (a typed verdict on a claim over five orthogonal axes — method, binding, verifier, agreement, authority — whose `ok`/`sad`/`bad` `projectedState` is a render-time projection, not a stored scalar) and `AssayStandard` (a verifier's measured, versioned reliability — the calibration reference every assay must point at). Includes canonical `ok`/`sad`/`bad` examples, ADR (`docs/adr/ADR-epistemic-assay-verdict-v0-1.md`), and a `validate-reasoning-examples` target that also enforces projection soundness (recomputes `assay()` from the stored axes and fails on drift) and retroactively brings the existing reasoning family under `make validate`.
+- `ReasoningReceipt.assay` (optional) — the reserved receipt landing spot for the Assay: a render-time, run-level epistemic summary that references its run's `ReasoningAssay` records by URN (`assayRefs`) and surfaces the projected `overallState`, weakest-link `binding`, verifier `calibrationRef`, per-state `counts`, and `projectedAt`. A cache of the referenced verdicts, not authoritative over them — re-projectable when an `AssayStandard` improves. Backward-compatible (optional field). The assay's `authority` axis continues to mirror `EventEnvelope.actor`/`integrity`, so the envelope needs no structural change.
 - Onboarding control-plane contract family: `WorkspaceScope`, `TrustMode`, `CapabilityPack`, `ConnectorActionScope`, `AutomationTemplate`, and `OnboardingReceipt`, with canonical first-run examples, semantic vocabulary seed, ADR, and `validate-onboarding-examples` validation target.
 - `ReasoningEvent.controlFlow` (optional) + a reserved control-flow `eventType` vocabulary (`reasoning.tool.called` / `reasoning.decision.branched` / `reasoning.subrun.spawned` / `reasoning.subrun.joined` / `reasoning.run.completed`) so emitters can carry a run's operational control flow for downstream narration-fidelity verification (SP-TRACE-CFR). Backward-compatible; operational structure only (no raw reasoning). See `docs/reasoning-control-flow-events.md` and `examples/reasoning_event_control_flow.json`.
 - SourceOS interaction substrate top-level index and README discovery links for `SourceOSInteractionEvent`, generated TypeScript/Python artifacts, and the Noetica → Superconscious → AgentPlane → AgentTerm reference flow.

examples/reasoning_receipt.json

@@ -12,6 +12,14 @@
     "memory": "proposal-only",
     "approval": "not-required"
   },
+  "assay": {
+    "assayRefs": ["urn:srcos:reasoning-assay:superconscious-demo-ok"],
+    "overallState": "ok",
+    "binding": "inline",
+    "calibrationRef": "urn:srcos:assay-standard:narration-fidelity-verifier:0.2.0",
+    "counts": { "ok": 1, "sad": 0, "bad": 0 },
+    "projectedAt": "2026-07-05T00:00:02Z"
+  },
   "replayClass": "exact",
   "capturedAt": "2026-05-05T00:00:01Z"
 }

schemas/ReasoningReceipt.json

@@ -18,6 +18,46 @@
       "type": "object",
       "additionalProperties": true
     },
+    "assay": {
+      "type": "object",
+      "additionalProperties": false,
+      "description": "Run-level epistemic posture: a render-time summary projected from the run's ReasoningAssay records. A cache, not the source of truth — re-derivable by re-projecting the referenced assays against the current AssayStandard, so it is optional and never authoritative over the assays themselves. The reserved receipt landing spot for binding + verifier calibration.",
+      "properties": {
+        "assayRefs": {
+          "type": "array",
+          "items": { "type": "string", "pattern": "^urn:srcos:reasoning-assay:" },
+          "description": "URNs of the ReasoningAssay records covering this run's claims. Referenced, not embedded, so each verdict stays independently re-projectable."
+        },
+        "overallState": {
+          "enum": ["ok", "sad", "bad"],
+          "description": "assay() projection for the run's overall answer AT projectedAt — the weakest-link readout across the referenced assays. A cache of the tuple, not a stored severity."
+        },
+        "binding": {
+          "enum": ["inline", "post-hoc"],
+          "description": "Weakest evidence binding across the run's assays. A single post-hoc claim pulls the run's binding to post-hoc; recorded so a post-hoc 'verified' run is not mistaken for an inline one."
+        },
+        "calibrationRef": {
+          "type": "string",
+          "pattern": "^urn:srcos:assay-standard:",
+          "description": "The AssayStandard the overall verdict rests on. An absent or uncalibrated reference caps overallState at 'sad'."
+        },
+        "counts": {
+          "type": "object",
+          "additionalProperties": false,
+          "description": "Tally of the referenced assays by projected state at projectedAt.",
+          "properties": {
+            "ok": { "type": "integer", "minimum": 0 },
+            "sad": { "type": "integer", "minimum": 0 },
+            "bad": { "type": "integer", "minimum": 0 }
+          }
+        },
+        "projectedAt": {
+          "type": "string",
+          "format": "date-time",
+          "description": "When this summary was projected. Re-projection against a newer AssayStandard yields a later summary rather than mutating this one."
+        }
+      }
+    },
     "replayClass": { "enum": ["exact", "best-effort", "evidence-only", "non-replayable-side-effect"] },
     "capturedAt": { "type": "string", "format": "date-time" }
   }

schemas/reasoning-contracts.README.md

@@ -54,6 +54,14 @@ historical assays re-project without mutating their records — the same discipl
 arm rather than only its current winner. `tools/validate_reasoning_examples.py` enforces this: it recomputes
 `assay()` from the stored axes and fails CI if the recorded `projectedState` does not follow from them.
 
+`ReasoningReceipt.assay` is the reserved receipt landing spot: a run-level, render-time summary that
+**references** its run's `ReasoningAssay` records by URN (`assayRefs`) rather than embedding them — so each
+verdict stays independently re-projectable — and surfaces the projected `overallState`, the weakest-link
+`binding`, the verifier `calibrationRef`, per-state `counts`, and `projectedAt`. It is a cache, never
+authoritative over the assays themselves, and optional so existing receipts remain valid. The `authority`
+axis of each assay mirrors `EventEnvelope.actor`/`integrity`, so authenticity is corroborated from the
+envelope the claim arrived on rather than asserted by the content — the envelope needs no structural change.
+
 Expected consumers:
 
 - `SocioProphet/superconscious` as the reference governed cognition loop;