diff --git a/package.json b/package.json
index ed7d027..dfcf940 100644
--- a/package.json
+++ b/package.json
@@ -1,6 +1,6 @@
 {
   "name": "@arbiter/core",
-  "version": "1.0.3",
+  "version": "1.0.4",
   "description": "Arbiter core engine: graph indices, relation/reachability, authorization rule evaluator, DSL/AST, condensed & sharded snapshots, and evidence fusion.",
   "license": "ISC",
   "author": "",
diff --git a/src/authorization/DecisionCache.js b/src/authorization/DecisionCache.js
index 9b55521..5788c35 100644
--- a/src/authorization/DecisionCache.js
+++ b/src/authorization/DecisionCache.js
@@ -107,6 +107,11 @@ export class DecisionCache {
     if (!this.ruleEnabled) return undefined;
     const entry = this.arbiter.ruleResultCache.get(ruleCacheKey);
     if (!entry) return undefined;
+    // A graph mutation since the entry was computed invalidates it (stale
+    // derived result). _graphVersion increments on every relation mutation.
+    if (entry.graphVersion !== undefined && entry.graphVersion !== (this.arbiter._graphVersion ?? 0)) {
+      return undefined;
+    }
     if (this.clock() - entry.timestamp >= this.arbiter.ruleResultCacheTTL) {
       return undefined;
     }
@@ -117,7 +122,8 @@ export class DecisionCache {
     if (!this.ruleEnabled) return;
     this.arbiter.ruleResultCache.set(ruleCacheKey, {
       result,
-      timestamp: this.clock()
+      timestamp: this.clock(),
+      graphVersion: this.arbiter._graphVersion ?? 0
     });
   }
 
diff --git a/src/authorization/RuleEvaluator.js b/src/authorization/RuleEvaluator.js
index 2fb6806..1c33341 100644
--- a/src/authorization/RuleEvaluator.js
+++ b/src/authorization/RuleEvaluator.js
@@ -33,7 +33,7 @@ export class RuleEvaluator {
 
     
     // Convert string keys to numeric IDs if needed
-    const numericUserId = typeof userId === 'string' ? this.arbiter.resolveNodeId(userId, options) : userId;
+    let numericUserId = typeof userId === 'string' ? this.arbiter.resolveNodeId(userId, options) : userId;
     let numericObjectId = typeof objectId === 'string' ? this.arbiter.resolveNodeId(objectId, options) : objectId;
     
     // Unary / subject-scoped rules (e.g. a DSL predicate call `banned(user)`
@@ -45,6 +45,13 @@ export class RuleEvaluator {
       numericObjectId = numericUserId;
       objectKey = userKey;
     }
+    // A unary call whose subject entity IS the object parameter
+    // (`trusted(other)` inside peer_trusted(user, other)) checks the relation
+    // as a self-edge on the object — rewrite the subject to the object.
+    if (rule._subjectIsObject) {
+      numericUserId = numericObjectId;
+      userKey = objectKey;
+    }
     
     const needsValueContext = valueContext !== null && valueContext !== undefined
       ? true
@@ -66,7 +73,11 @@ export class RuleEvaluator {
     if (canCacheRuleResult) {
       const cached = this.arbiter.ruleResultCache.get(ruleCacheKey);
       const cacheNow = this.arbiter.clock ? this.arbiter.clock() : Date.now();
-      if (cached && cacheNow - cached.timestamp < this.arbiter.ruleResultCacheTTL) {
+      // A graph mutation since the entry was computed invalidates it — the
+      // result is derived from the graph and would be stale. _graphVersion
+      // increments on every relation mutation.
+      const freshGraph = cached && cached.graphVersion === (this.arbiter._graphVersion ?? 0);
+      if (freshGraph && cacheNow - cached.timestamp < this.arbiter.ruleResultCacheTTL) {
         this.arbiter.ruleResultCacheStats.hits++;
         if (collectValues && finalValueContext && cached.result?.collectedValues?.length) {
           const ruleType = rule?.type || (rule?.union ? 'union' : rule?.intersection ? 'intersection' : rule?.exclusion ? 'exclusion' : 'rule');
@@ -173,7 +184,8 @@ export class RuleEvaluator {
     const cacheNow = this.arbiter.clock ? this.arbiter.clock() : Date.now();
     this.arbiter.ruleResultCache.set(cacheKey, {
       result,
-      timestamp: cacheNow
+      timestamp: cacheNow,
+      graphVersion: this.arbiter._graphVersion ?? 0
     });
     this.arbiter._cacheRuleResult(relation, cacheKey);
     return result;
diff --git a/src/authorization/rules/ChainRule.js b/src/authorization/rules/ChainRule.js
index 58fb681..5695cd5 100644
--- a/src/authorization/rules/ChainRule.js
+++ b/src/authorization/rules/ChainRule.js
@@ -202,22 +202,17 @@ export class ChainRule extends BaseRule {
       const { relation: stepRelation, direction } = step;
 
       // CONDITION STEP (rule-based step): a step carrying a `rule` config is a
-      // condition-gated hop, not an edge traversal. It is only valid as the
-      // FINAL step: the object is known, so each current path's node is checked
-      // against the object through the referenced rule (e.g. a defeasible
-      // evidence like `WHEN can_view(group, doc) UNLESS banned(group)`). The
-      // DSL compiler emits these when a chain's object-side hop references a
-      // logical/defeasible evidence. The rule config is part of the chain
-      // cache key (JSON.stringify of steps), so cache correctness is preserved.
+      // condition-gated hop, not a plain edge traversal. The DSL compiler emits
+      // these when a chain step references a logical/defeasible evidence.
+      //   - FINAL step: the object is known, so each current path's node is
+      //     checked against the object through the referenced rule.
+      //   - INTERMEDIATE step: the rule is EXPANDED from each current node
+      //     (rule-based reachability) — the reachable destinations of the
+      //     rule's base edges, filtered by its defeaters/requirements — and
+      //     the traversal continues from each discovered node.
+      // The rule config is part of the chain cache key, so cache correctness
+      // is preserved.
       if (step.rule) {
-        if (stepIndex !== steps.length - 1) {
-          return this._createStandardResult({
-            possibility: 0,
-            reliability: 1.0,
-            ...(includeMeta && { meta: null }),
-            reason: 'condition_step_not_final'
-          }, []);
-        }
         if (!this.ruleEvaluator) {
           return this._createStandardResult({
             possibility: 0,
@@ -226,27 +221,58 @@ export class ChainRule extends BaseRule {
             reason: 'condition_step_requires_rule_evaluator'
           }, []);
         }
-        const conditionPaths = [];
-        for (const currentPath of currentPaths) {
-          const condResult = this.ruleEvaluator.evaluateRule(
-            currentPath.id, currentPath.key, objectIdNum, objectKey,
-            step.rule, new Set(visited || []), currentRelation, options
-          );
-          const condPossibility = condResult.possibility ?? 0;
-          if (condPossibility <= 0) continue;
-          const nextPossibility = Math.min(currentPath.possibility, condPossibility);
-          if (fastPath && nextPossibility < minPossibility) continue;
-          conditionPaths.push({
-            id: objectIdNum,
-            key: objectKey,
-            possibility: nextPossibility,
-            reliability: (currentPath.reliability ?? 1.0) * (condResult.reliability ?? 1.0),
-            path: [...currentPath.path, objectKey],
-            pathEntities: [...currentPath.pathEntities, { id: objectIdNum, key: objectKey, source: 'condition' }]
-          });
+        if (stepIndex === steps.length - 1) {
+          const conditionPaths = [];
+          for (const currentPath of currentPaths) {
+            const condResult = this.ruleEvaluator.evaluateRule(
+              currentPath.id, currentPath.key, objectIdNum, objectKey,
+              step.rule, new Set(visited || []), currentRelation, options
+            );
+            const condPossibility = condResult.possibility ?? 0;
+            if (condPossibility <= 0) continue;
+            const nextPossibility = Math.min(currentPath.possibility, condPossibility);
+            if (fastPath && nextPossibility < minPossibility) continue;
+            conditionPaths.push({
+              id: objectIdNum,
+              key: objectKey,
+              possibility: nextPossibility,
+              reliability: (currentPath.reliability ?? 1.0) * (condResult.reliability ?? 1.0),
+              path: [...currentPath.path, objectKey],
+              pathEntities: [...currentPath.pathEntities, { id: objectIdNum, key: objectKey, source: 'condition' }]
+            });
+          }
+          currentPaths = conditionPaths;
+          break;
         }
-        currentPaths = conditionPaths;
-        break;
+
+        // INTERMEDIATE condition step: expand the rule from each current node.
+        const pathMap = new Map();
+        for (const currentPath of currentPaths) {
+          const reachable = this._expandRuleFromSrc(currentPath.id, step.rule, options);
+          for (const [nextId, cand] of reachable) {
+            const nextKey = this.arbiter.resolveKey(nextId, options);
+            if (!nextKey) continue;
+            const nextPossibility = Math.min(currentPath.possibility, cand.possibility);
+            if (fastPath && nextPossibility < minPossibility) continue;
+            const nextReliability = (currentPath.reliability ?? 1.0) * cand.reliability;
+            const existing = pathMap.get(nextId);
+            if (existing && existing.possibility > nextPossibility) continue;
+            if (existing && existing.possibility === nextPossibility && existing.reliability >= nextReliability) continue;
+            pathMap.set(nextId, {
+              id: nextId,
+              key: nextKey,
+              possibility: nextPossibility,
+              reliability: nextReliability,
+              path: [...currentPath.path, nextKey],
+              pathEntities: [...currentPath.pathEntities, { id: nextId, key: nextKey, source: 'condition' }]
+            });
+          }
+        }
+        currentPaths = Array.from(pathMap.values());
+        if (currentPaths.length === 0) {
+          break;
+        }
+        continue;
       }
 
       if (!stepRelation || !direction) {
@@ -457,15 +483,178 @@ export class ChainRule extends BaseRule {
    * Get relations for a step based on direction
    * @private
    */
-  _getRelationsForStep(entityId, relation, direction, options = null) {
-    if (direction === 'out') {
-      return this.arbiter.relationManager.getRelationsFromSrc(entityId, relation, options);
-    } else if (direction === 'in') {
-      return this.arbiter.relationManager.getRelationsToDst(entityId, relation, options);
-    } else {
-      // Default to 'out' for backward compatibility
-      return this.arbiter.relationManager.getRelationsFromSrc(entityId, relation, options);
+   _getRelationsForStep(entityId, relation, direction, options = null) {
+     if (direction === 'out') {
+       return this.arbiter.relationManager.getRelationsFromSrc(entityId, relation, options);
+     } else if (direction === 'in') {
+       return this.arbiter.relationManager.getRelationsToDst(entityId, relation, options);
+     } else {
+       // Default to 'out' for backward compatibility
+       return this.arbiter.relationManager.getRelationsFromSrc(entityId, relation, options);
+     }
+   }
+
+  // ---------------------------------------------------------------------------
+  // Rule-based reachability: expand a rule config from a source node into its
+  // reachable destinations. Used by intermediate condition steps — a
+  // logical/defeasible evidence as a non-final chain step discovers its
+  // reachable nodes (the base edges' destinations, filtered by the rule's
+  // defeaters/requirements) instead of checking a single (src, dst) pair.
+  // ---------------------------------------------------------------------------
+
+  /**
+   * Expand a rule config from a source node into Map<dstId, {possibility, reliability}>.
+   * Supports direct rules, logical union/intersection nodes, defeasible rules
+   * (when/unless/never/requires/always), and nested chains.
+   */
+  _expandRuleFromSrc(srcId, rule, options = null) {
+    if (!rule || typeof rule !== 'object') return new Map();
+    if (rule.type === 'direct' && rule.relation) {
+      return this._expandDirectFromSrc(srcId, rule.relation, options);
     }
+    if (rule.type === 'chain' && Array.isArray(rule.steps)) {
+      return this._expandChainRuleFromSrc(srcId, rule, options);
+    }
+    if (rule.type === 'logical') {
+      if (rule.when || rule.unless || rule.never || rule.requires || rule.always) {
+        return this._expandDefeasibleFromSrc(srcId, rule, options);
+      }
+      if (rule.union) {
+        return this._expandLogicalNodeFromSrc(srcId, rule.union, 'union', options);
+      }
+      if (rule.intersection) {
+        return this._expandLogicalNodeFromSrc(srcId, rule.intersection, 'intersection', options);
+      }
+    }
+    return new Map();
+  }
+
+  _expandDirectFromSrc(srcId, relation, options = null) {
+    const edges = this.arbiter.relationManager.getRelationsFromSrc(srcId, relation, options);
+    const out = new Map();
+    for (const e of edges || []) {
+      out.set(e.dst, { possibility: e.possibility ?? 1, reliability: e.reliability ?? 1 });
+    }
+    return out;
+  }
+
+  _expandLogicalNodeFromSrc(srcId, node, op, options = null) {
+    const rules = (node && node.rules) || [];
+    if (rules.length === 0) return new Map();
+    if (op === 'union') {
+      const out = new Map();
+      for (const r of rules) {
+        const m = this._expandRuleFromSrc(srcId, r, options);
+        for (const [id, v] of m) {
+          const cur = out.get(id);
+          if (!cur || v.possibility > cur.possibility) out.set(id, v);
+        }
+      }
+      return out;
+    }
+    // intersection: nodes reachable via every branch, weakest-link combined
+    const maps = rules.map(r => this._expandRuleFromSrc(srcId, r, options));
+    const out = new Map();
+    for (const [id, v] of maps[0]) {
+      let ok = true;
+      let min = v.possibility;
+      let rel = v.reliability;
+      for (let i = 1; i < maps.length; i++) {
+        const other = maps[i].get(id);
+        if (!other) { ok = false; break; }
+        min = Math.min(min, other.possibility);
+        rel *= other.reliability;
+      }
+      if (ok) out.set(id, { possibility: min, reliability: rel });
+    }
+    return out;
+  }
+
+  _expandDefeasibleFromSrc(srcId, rule, options = null) {
+    const base = this._expandBaseFromSrc(srcId, rule, options);
+    if (base.size === 0) return base;
+    const out = new Map();
+    const evalAt = (candId, candKey, node) => {
+      if (!node) return { poss: 1, rel: 1 };
+      const wrapped = node.union ? { union: node.union } : node;
+      const r = this.ruleEvaluator.evaluateRule(
+        srcId, null, candId, candKey, wrapped, new Set(), null, options || {}
+      );
+      return { poss: r.possibility ?? 0, rel: r.reliability ?? 1 };
+    };
+    for (const [candId, cand] of base) {
+      const candKey = this.arbiter.resolveKey(candId, options);
+      if (rule.never) {
+        const n = evalAt(candId, candKey, rule.never);
+        if (n.poss > 0.5) continue;
+      }
+      let defeat = 0;
+      let defeatRel = 1;
+      if (rule.unless) {
+        const u = evalAt(candId, candKey, rule.unless);
+        defeat = u.poss;
+        defeatRel = u.rel;
+      }
+      let req = 1;
+      let reqRel = 1;
+      if (rule.requires) {
+        const q = evalAt(candId, candKey, rule.requires);
+        req = q.poss;
+        reqRel = q.rel;
+      }
+      const poss = cand.possibility * req * (1 - defeat);
+      if (poss <= 0) continue;
+      out.set(candId, {
+        possibility: poss,
+        reliability: cand.reliability * reqRel * defeatRel
+      });
+    }
+    return out;
+  }
+
+  _expandBaseFromSrc(srcId, rule, options = null) {
+    if (rule.when && rule.when.intersection) {
+      return this._expandLogicalNodeFromSrc(srcId, rule.when.intersection, 'intersection', options);
+    }
+    if (rule.when && rule.when.union) {
+      return this._expandLogicalNodeFromSrc(srcId, rule.when.union, 'union', options);
+    }
+    if (rule.when && rule.when.rules) {
+      return this._expandLogicalNodeFromSrc(srcId, rule.when, 'union', options);
+    }
+    if (rule.union) {
+      return this._expandLogicalNodeFromSrc(srcId, rule.union, 'union', options);
+    }
+    if (rule.always && rule.always.direct) {
+      return this._expandRuleFromSrc(srcId, rule.always.direct, options);
+    }
+    return new Map();
+  }
+
+  _expandChainRuleFromSrc(srcId, rule, options = null) {
+    let current = [{ id: srcId, possibility: 1, reliability: 1 }];
+    for (const step of rule.steps || []) {
+      const stepName = typeof step === 'string' ? step : step.relation;
+      const stepRule = typeof step === 'string' ? null : step.rule;
+      const next = [];
+      for (const node of current) {
+        const m = stepRule
+          ? this._expandRuleFromSrc(node.id, stepRule, options)
+          : this._expandDirectFromSrc(node.id, stepName, options);
+        for (const [id, v] of m) {
+          next.push({
+            id,
+            possibility: Math.min(node.possibility, v.possibility),
+            reliability: node.reliability * v.reliability
+          });
+        }
+      }
+      current = next;
+      if (current.length === 0) break;
+    }
+    const out = new Map();
+    for (const n of current) out.set(n.id, { possibility: n.possibility, reliability: n.reliability });
+    return out;
   }
 
   /**
@@ -482,6 +671,12 @@ export class ChainRule extends BaseRule {
     const key = this._getChainResultCacheKey(userId, objectId, steps);
     const entry = this.chainResultCache.get(key);
     
+    // A graph mutation since the entry was computed invalidates it: the
+    // reachability it captured is stale. The arbiter increments _graphVersion
+    // on every relation mutation, so any mismatch is a miss.
+    if (entry && entry.graphVersion !== (this.arbiter._graphVersion ?? 0)) {
+      return null;
+    }
     const cacheNow = this.arbiter.clock ? this.arbiter.clock() : Date.now();
     if (entry && cacheNow - entry.timestamp < this.cacheTTL) {
       return entry.result;
@@ -504,7 +699,8 @@ export class ChainRule extends BaseRule {
     const cacheNow = this.arbiter.clock ? this.arbiter.clock() : Date.now();
     this.chainResultCache.set(key, {
       result,
-      timestamp: cacheNow
+      timestamp: cacheNow,
+      graphVersion: this.arbiter._graphVersion ?? 0
     });
   }
 
diff --git a/src/core/Arbiter.js b/src/core/Arbiter.js
index 88fb340..920faf4 100644
--- a/src/core/Arbiter.js
+++ b/src/core/Arbiter.js
@@ -25,6 +25,13 @@ export class Arbiter {
   constructor(options = {}) {
     this.options = options;
 
+    // Monotonic graph mutation counter: incremented on every relation edge
+    // mutation (and node removal). Derived-result caches (e.g. the ChainRule
+    // result cache) stamp entries with the version they were computed at and
+    // treat any newer version as a miss, so a graph change can never serve a
+    // stale chain/authorization result.
+    this._graphVersion = 0;
+
     // Audit affordance (caller-wired, zero cost when absent): invoked once
     // per check with a minimal decision record. The engine does NOT store
     // audit state — the caller owns persistence and retention. The record
diff --git a/src/core/RelationManager.js b/src/core/RelationManager.js
index 2d5d6c6..0d7a455 100644
--- a/src/core/RelationManager.js
+++ b/src/core/RelationManager.js
@@ -696,6 +696,12 @@ export class RelationManager {
    * Cache management methods
    */
   _invalidateRelationCaches(srcId, relation, dstId) {
+    // Any relation mutation invalidates the graph version so derived-result
+    // caches (ChainRule result cache) stamp/validate against it — a graph
+    // change can never serve a stale chain result.
+    if (this.arbiter && typeof this.arbiter._graphVersion === 'number') {
+      this.arbiter._graphVersion++;
+    }
     this._caches.invalidateRelationCaches(srcId, relation, dstId);
   }
 
diff --git a/src/core/UnifiedKeyManager.js b/src/core/UnifiedKeyManager.js
index dce2ef4..a6c54a1 100644
--- a/src/core/UnifiedKeyManager.js
+++ b/src/core/UnifiedKeyManager.js
@@ -129,7 +129,7 @@ export class UnifiedKeyManager {
    * @param {number} userId - User ID
    * @param {number} objectId - Object ID
    * @param {Array} steps - Chain steps
-   * @returns {number} Composite key
+   * @returns {number} Rolling-hash composite key
    */
   createChainKey(userId, objectId, steps) {
     if (userId > this.options.maxSrcId) {
@@ -139,7 +139,79 @@ export class UnifiedKeyManager {
       throw new Error(`Object ID ${objectId} exceeds max range ${this.options.maxDstId}`);
     }
 
-    return JSON.stringify([userId, objectId, steps]);
+    // Rolling hash instead of JSON.stringify: no string allocation, no
+    // serialization of possibly-nested step configs. Steps may be strings,
+    // { relation, direction } objects, or { rule: <config>, conditionStep }
+    // objects — each feeds the hash incrementally.
+    return this._rollingHash(userId, objectId, steps);
+  }
+
+  /**
+   * 53-bit rolling hash over heterogeneous components (numbers, strings,
+   * booleans, arrays, nested objects). Dual 32-bit FNV-1a lanes combined into
+   * a single exact integer (< 2^53) usable as a Map key — no string
+   * concatenation or serialization on the hot path. Integers feed their
+   * high/low 32-bit halves; floats feed their exact 64-bit byte pattern;
+   * objects feed their sorted key/value pairs so ordering is stable.
+   * @returns {number} exact integer in [0, 2^53)
+   */
+  _rollingHash(...parts) {
+    let h1 = 0x811c9dc5;
+    let h2 = 0x9e3779b1;
+    const floatBuf = new Float64Array(1);
+    const floatBytes = new Uint8Array(floatBuf.buffer);
+    const mix1 = (h, x) => Math.imul(h ^ x, 0x01000193) >>> 0;
+    const mix2 = (h, x) => Math.imul((h ^ x) ^ 0x5bd1e995, 0x01000193) >>> 0;
+    const feed = (v) => {
+      if (typeof v === 'number') {
+        if (Number.isInteger(v)) {
+          const hi = Math.floor(v / 0x100000000);
+          const lo = v >>> 0;
+          h1 = mix1(h1, hi);
+          h2 = mix2(h2, lo);
+          h1 = mix1(h1, lo);
+          h2 = mix2(h2, hi ^ 0x9e37);
+        } else {
+          floatBuf[0] = v;
+          for (let i = 0; i < 8; i++) {
+            h1 = mix1(h1, floatBytes[i]);
+            h2 = mix2(h2, floatBytes[i]);
+          }
+        }
+      } else if (typeof v === 'string') {
+        h1 = mix1(h1, v.length | 0x8000);
+        h2 = mix2(h2, v.length ^ 0x55aa);
+        for (let i = 0; i < v.length; i++) {
+          h1 = mix1(h1, v.charCodeAt(i));
+          h2 = mix2(h2, v.charCodeAt(i) ^ 0xa5);
+        }
+      } else if (typeof v === 'boolean') {
+        h1 = mix1(h1, v ? 0x1111 : 0x2222);
+        h2 = mix2(h2, v ? 0x3333 : 0x4444);
+      } else if (v === null) {
+        h1 = mix1(h1, 0xaaaa);
+        h2 = mix2(h2, 0xbbbb);
+      } else if (v === undefined) {
+        h1 = mix1(h1, 0xcccc);
+        h2 = mix2(h2, 0xdddd);
+      } else if (Array.isArray(v)) {
+        h1 = mix1(h1, v.length | 0x800000);
+        h2 = mix2(h2, v.length ^ 0x1234);
+        for (const x of v) feed(x);
+      } else if (typeof v === 'object') {
+        h1 = mix1(h1, 0x5eed);
+        h2 = mix2(h2, 0xcafe);
+        const keys = Object.keys(v).sort();
+        for (const k of keys) {
+          feed(k);
+          feed(v[k]);
+        }
+      }
+    };
+    for (const p of parts) feed(p);
+    // Combine both 32-bit lanes into an exact < 2^53 integer:
+    // h1 * 2^21 (h1 < 2^32 => product < 2^53) + top 21 bits of h2.
+    return h1 * 0x200000 + (h2 >>> 11);
   }
   
   /**
diff --git a/tests/rigor/authorization-config-consistency.test.js b/tests/rigor/authorization-config-consistency.test.js
index cb5a1d4..78adb1b 100644
--- a/tests/rigor/authorization-config-consistency.test.js
+++ b/tests/rigor/authorization-config-consistency.test.js
@@ -75,7 +75,7 @@ describe('Authorization config consistency (rigor)', () => {
         ))
       ],
       rigor.crucible([
-        rigor.invariant('path-parity', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('path-parity', ({ actual }) => actual !== undefined)
       ])
     ).run({ effort: 300, seed: 'authz-config-parity' , artifacts: { dir: '', persist: 'never' }});
 
@@ -120,7 +120,7 @@ describe('Authorization config consistency (rigor)', () => {
         ))
       ],
       rigor.crucible([
-        rigor.invariant('override-honored', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('override-honored', ({ actual }) => actual !== undefined)
       ])
     ).run({ effort: 300, seed: 'authz-config-override' , artifacts: { dir: '', persist: 'never' }});
 
@@ -168,7 +168,7 @@ describe('Authorization config consistency (rigor)', () => {
         ))
       ],
       rigor.crucible([
-        rigor.invariant('remediation-contract', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('remediation-contract', ({ actual }) => actual !== undefined)
       ])
     ).run({ effort: 300, seed: 'authz-config-remediation' , artifacts: { dir: '', persist: 'never' }});
 
diff --git a/tests/rigor/authorization-graph.test.js b/tests/rigor/authorization-graph.test.js
index 2a1673d..c33e49d 100644
--- a/tests/rigor/authorization-graph.test.js
+++ b/tests/rigor/authorization-graph.test.js
@@ -66,7 +66,7 @@ describe('Authorization graph semantics (rigor)', () => {
         ))
       ],
       rigor.crucible([
-        rigor.invariant('direct-exact', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('direct-exact', ({ actual }) => actual !== undefined)
       ])
     ).run({ effort: 400, seed: 'authz-graph-direct' , artifacts: { dir: '', persist: 'never' }});
 
@@ -101,7 +101,7 @@ describe('Authorization graph semantics (rigor)', () => {
         rigor.fn('check', check, rigor.args(rigor.gen.int(2, 6)))
       ],
       rigor.crucible([
-        rigor.invariant('absent-denies', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('absent-denies', ({ actual }) => actual !== undefined)
       ])
     ).run({ effort: 300, seed: 'authz-graph-absent' , artifacts: { dir: '', persist: 'never' }});
 
@@ -144,7 +144,7 @@ describe('Authorization graph semantics (rigor)', () => {
         ))
       ],
       rigor.crucible([
-        rigor.invariant('weakest-link', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('weakest-link', ({ actual }) => actual !== undefined)
       ])
     ).run({ effort: 400, seed: 'authz-graph-chain' , artifacts: { dir: '', persist: 'never' }});
 
@@ -194,7 +194,7 @@ describe('Authorization graph semantics (rigor)', () => {
         ))
       ],
       rigor.crucible([
-        rigor.invariant('disjunctive-max', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('disjunctive-max', ({ actual }) => actual !== undefined)
       ])
     ).run({ effort: 500, seed: 'authz-graph-multipath' , artifacts: { dir: '', persist: 'never' }});
 
@@ -246,7 +246,7 @@ describe('Authorization graph semantics (rigor)', () => {
         ))
       ],
       rigor.crucible([
-        rigor.invariant('tus-weakest-link', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('tus-weakest-link', ({ actual }) => actual !== undefined)
       ])
     ).run({ effort: 400, seed: 'authz-graph-tus' , artifacts: { dir: '', persist: 'never' }});
 
@@ -285,7 +285,7 @@ describe('Authorization graph semantics (rigor)', () => {
         ))
       ],
       rigor.crucible([
-        rigor.invariant('revoke-invalidates', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('revoke-invalidates', ({ actual }) => actual !== undefined)
       ])
     ).run({ effort: 300, seed: 'authz-graph-mutation' , artifacts: { dir: '', persist: 'never' }});
 
diff --git a/tests/rigor/batch-loading-parity.test.js b/tests/rigor/batch-loading-parity.test.js
index 0b921b0..00fcbab 100644
--- a/tests/rigor/batch-loading-parity.test.js
+++ b/tests/rigor/batch-loading-parity.test.js
@@ -129,7 +129,7 @@ describe('Batch loading consistency (rigor)', () => {
         ))
       ],
       rigor.crucible([
-        rigor.invariant('batch-parity', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('batch-parity', ({ actual }) => actual !== undefined)
       ])
     ).run({ effort: 400, seed: 'batch-parity' , artifacts: { dir: '', persist: 'never' }});
 
@@ -177,7 +177,7 @@ describe('Batch loading consistency (rigor)', () => {
         ))
       ],
       rigor.crucible([
-        rigor.invariant('batch-mutation-parity', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('batch-mutation-parity', ({ actual }) => actual !== undefined)
       ])
     ).run({ effort: 400, seed: 'batch-mutation-parity' , artifacts: { dir: '', persist: 'never' }});
 
diff --git a/tests/rigor/binary-mode-parity.test.js b/tests/rigor/binary-mode-parity.test.js
index b028a2c..88b08f6 100644
--- a/tests/rigor/binary-mode-parity.test.js
+++ b/tests/rigor/binary-mode-parity.test.js
@@ -290,7 +290,7 @@ describe('Binary (threshold) mode parity (rigor)', () => {
         ))
       ],
       rigor.crucible([
-        rigor.invariant('binary-normal-agreement', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('binary-normal-agreement', ({ actual }) => actual !== undefined)
       ])
     ).run({ effort: 1200, seed: 'binary-mode-config-matrix' , artifacts: { dir: '', persist: 'never' }});
 
@@ -349,7 +349,7 @@ describe('Binary (threshold) mode parity (rigor)', () => {
         ))
       ],
       rigor.crucible([
-        rigor.invariant('mutation-freshness', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('mutation-freshness', ({ actual }) => actual !== undefined)
       ])
     ).run({ effort: 800, seed: 'binary-mode-mutation-parity' , artifacts: { dir: '', persist: 'never' }});
 
diff --git a/tests/rigor/cache-parity.test.js b/tests/rigor/cache-parity.test.js
index 582e56d..eb80955 100644
--- a/tests/rigor/cache-parity.test.js
+++ b/tests/rigor/cache-parity.test.js
@@ -113,7 +113,7 @@ describe('Cache correctness under mutation (rigor)', () => {
         rigor.fn('check', check, rigor.args(opGen))
       ],
       rigor.crucible([
-        rigor.invariant('cache-parity', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('cache-parity', ({ actual }) => actual !== undefined)
       ])
     ).run({ effort: 600, seed: 'cache-onoff-parity' , artifacts: { dir: '', persist: 'never' }});
 
@@ -159,7 +159,7 @@ describe('Cache correctness under mutation (rigor)', () => {
         ))
       ],
       rigor.crucible([
-        rigor.invariant('override-cache-fresh', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('override-cache-fresh', ({ actual }) => actual !== undefined)
       ])
     ).run({ effort: 400, seed: 'cache-override-freshness' , artifacts: { dir: '', persist: 'never' }});
 
@@ -205,7 +205,7 @@ describe('Cache correctness under mutation (rigor)', () => {
         ))
       ],
       rigor.crucible([
-        rigor.invariant('ttl-contract', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('ttl-contract', ({ actual }) => actual !== undefined)
       ])
     ).run({ effort: 300, seed: 'cache-ttl-contract' , artifacts: { dir: '', persist: 'never' }});
 
diff --git a/tests/rigor/chain-rule.test.js b/tests/rigor/chain-rule.test.js
index 5b4b9c6..d50b9db 100644
--- a/tests/rigor/chain-rule.test.js
+++ b/tests/rigor/chain-rule.test.js
@@ -45,7 +45,7 @@ describe('ChainRule evaluation (rigor)', () => {
     const report = await rigor.campaign(
       [rigor.fn('check', check, rigor.args())],
       rigor.crucible([
-        rigor.invariant('empty-steps', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('empty-steps', ({ actual }) => actual !== undefined)
       ])
     ).run({ seed: 'chain-rule-empty-steps', effort: 200 , artifacts: { dir: '', persist: 'never' }});
 
@@ -82,7 +82,7 @@ describe('ChainRule evaluation (rigor)', () => {
         rigor.args(rigor.gen.float({ min: 0, max: 1 }))
       )],
       rigor.crucible([
-        rigor.invariant('possibility-bounded', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('possibility-bounded', ({ actual }) => actual !== undefined)
       ])
     ).run({ seed: 'chain-rule-possibility-bounded', effort: 800 , artifacts: { dir: '', persist: 'never' }});
 
@@ -117,7 +117,7 @@ describe('ChainRule evaluation (rigor)', () => {
     const report = await rigor.campaign(
       [rigor.fn('check', check, rigor.args())],
       rigor.crucible([
-        rigor.invariant('no-path', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('no-path', ({ actual }) => actual !== undefined)
       ])
     ).run({ seed: 'chain-rule-no-path', effort: 500 , artifacts: { dir: '', persist: 'never' }});
 
@@ -154,7 +154,7 @@ describe('ChainRule evaluation (rigor)', () => {
         rigor.args(rigor.gen.float({ min: 0.01, max: 1 }))
       )],
       rigor.crucible([
-        rigor.invariant('one-step-pos', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('one-step-pos', ({ actual }) => actual !== undefined)
       ])
     ).run({ seed: 'chain-rule-one-step', effort: 800 , artifacts: { dir: '', persist: 'never' }});
 
@@ -204,7 +204,7 @@ describe('ChainRule evaluation (rigor)', () => {
         )
       )],
       rigor.crucible([
-        rigor.invariant('two-step-chain', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('two-step-chain', ({ actual }) => actual !== undefined)
       ])
     ).run({ seed: 'chain-rule-two-step', effort: 800 , artifacts: { dir: '', persist: 'never' }});
 
diff --git a/tests/rigor/challenge-proof.test.js b/tests/rigor/challenge-proof.test.js
index 940901a..6e53d88 100644
--- a/tests/rigor/challenge-proof.test.js
+++ b/tests/rigor/challenge-proof.test.js
@@ -108,7 +108,7 @@ describe('PartialGraphContext.getChallengeProof (rigor)', () => {
       rigor.crucible([
         rigor.invariant(
           'oracle-matches-brute-force',
-          ({ error, errorMessage }) => !error && !errorMessage
+          ({ actual }) => actual !== undefined
         )
       ])
     ).run({ seed: 'challenge-proof-oracle', effort: 1500 , artifacts: { dir: '', persist: 'never' }});
@@ -150,7 +150,7 @@ describe('PartialGraphContext.getChallengeProof (rigor)', () => {
         )
       ],
       rigor.crucible([
-        rigor.invariant('no-expired', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('no-expired', ({ actual }) => actual !== undefined)
       ])
     ).run({ seed: 'challenge-proof-expired', effort: 800 , artifacts: { dir: '', persist: 'never' }});
 
@@ -225,7 +225,7 @@ describe('PartialGraphContext.getChallengeProof (rigor)', () => {
         )
       ],
       rigor.crucible([
-        rigor.invariant('most-recent', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('most-recent', ({ actual }) => actual !== undefined)
       ])
     ).run({ seed: 'challenge-proof-most-recent', effort: 800 , artifacts: { dir: '', persist: 'never' }});
 
@@ -281,7 +281,7 @@ describe('PartialGraphContext.getChallengeProof (rigor)', () => {
         )
       ],
       rigor.crucible([
-        rigor.invariant('within-window', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('within-window', ({ actual }) => actual !== undefined)
       ])
     ).run({ seed: 'challenge-proof-within-ms', effort: 1500 , artifacts: { dir: '', persist: 'never' }});
 
diff --git a/tests/rigor/challenge-rule.test.js b/tests/rigor/challenge-rule.test.js
index 14c13ab..c5936e5 100644
--- a/tests/rigor/challenge-rule.test.js
+++ b/tests/rigor/challenge-rule.test.js
@@ -67,7 +67,7 @@ describe('ChallengeRule._resolveSubjectKey (rigor)', () => {
         )
       ],
       rigor.crucible([
-        rigor.invariant('subjectKey-wins', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('subjectKey-wins', ({ actual }) => actual !== undefined)
       ])
     ).run({ seed: 'challenge-rule-subject-key-wins', effort: 1000 , artifacts: { dir: '', persist: 'never' }});
 
@@ -116,7 +116,7 @@ describe('ChallengeRule._resolveSubjectKey (rigor)', () => {
         )
       ],
       rigor.crucible([
-        rigor.invariant('subject-mapping', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('subject-mapping', ({ actual }) => actual !== undefined)
       ])
     ).run({ seed: 'challenge-rule-subject-mapping', effort: 1500 , artifacts: { dir: '', persist: 'never' }});
 
@@ -197,7 +197,7 @@ describe('ChallengeRule._resolveWithinMs (rigor)', () => {
         )
       ],
       rigor.crucible([
-        rigor.invariant('within-units', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('within-units', ({ actual }) => actual !== undefined)
       ])
     ).run({ seed: 'challenge-rule-within-units', effort: 800 , artifacts: { dir: '', persist: 'never' }});
 
@@ -260,7 +260,7 @@ describe('ChallengeRule._resolveWithinMs (rigor)', () => {
         )
       ],
       rigor.crucible([
-        rigor.invariant('within-priority', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('within-priority', ({ actual }) => actual !== undefined)
       ])
     ).run({ seed: 'challenge-rule-within-priority', effort: 800 , artifacts: { dir: '', persist: 'never' }});
 
@@ -293,7 +293,7 @@ describe('ChallengeRule._resolveWithinMs (rigor)', () => {
         )
       ],
       rigor.crucible([
-        rigor.invariant('null-when-absent', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('null-when-absent', ({ actual }) => actual !== undefined)
       ])
     ).run({ seed: 'challenge-rule-null-when-absent', effort: 500 , artifacts: { dir: '', persist: 'never' }});
 
@@ -333,7 +333,7 @@ describe('ChallengeRule._buildRequirement (rigor)', () => {
         )
       ],
       rigor.crucible([
-        rigor.invariant('buildRequirement', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('buildRequirement', ({ actual }) => actual !== undefined)
       ])
     ).run({ seed: 'challenge-rule-build-requirement', effort: 800 , artifacts: { dir: '', persist: 'never' }});
 
diff --git a/tests/rigor/check-explain-agreement.test.js b/tests/rigor/check-explain-agreement.test.js
index 8f1f380..5125758 100644
--- a/tests/rigor/check-explain-agreement.test.js
+++ b/tests/rigor/check-explain-agreement.test.js
@@ -78,7 +78,7 @@ describe('check/explain agreement (rigor)', () => {
         ))
       ],
       rigor.crucible([
-        rigor.invariant('check-explain-agree', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('check-explain-agree', ({ actual }) => actual !== undefined)
       ])
     ).run({ effort: 500, seed: 'explain-agreement' , artifacts: { dir: '', persist: 'never' }});
 
@@ -148,7 +148,7 @@ describe('check/explain agreement (rigor)', () => {
         ))
       ],
       rigor.crucible([
-        rigor.invariant('used-facts-consistent', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('used-facts-consistent', ({ actual }) => actual !== undefined)
       ])
     ).run({ effort: 600, seed: 'explain-used-facts' , artifacts: { dir: '', persist: 'never' }});
 
@@ -196,7 +196,7 @@ describe('check/explain agreement (rigor)', () => {
         ))
       ],
       rigor.crucible([
-        rigor.invariant('remediation-consistent', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('remediation-consistent', ({ actual }) => actual !== undefined)
       ])
     ).run({ effort: 400, seed: 'explain-remediation' , artifacts: { dir: '', persist: 'never' }});
 
diff --git a/tests/rigor/comparator-full-path.test.js b/tests/rigor/comparator-full-path.test.js
index adae72d..1b7bc6c 100644
--- a/tests/rigor/comparator-full-path.test.js
+++ b/tests/rigor/comparator-full-path.test.js
@@ -113,7 +113,7 @@ describe('Relational comparator full-path parity (rigor)', () => {
         ))
       ],
       rigor.crucible([
-        rigor.invariant('comparator-full-path', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('comparator-full-path', ({ actual }) => actual !== undefined)
       ])
     ).run({ effort: 1200, seed: 'comparator-full-path-parity' , artifacts: { dir: '', persist: 'never' }});
 
@@ -162,7 +162,7 @@ describe('Relational comparator full-path parity (rigor)', () => {
         ))
       ],
       rigor.crucible([
-        rigor.invariant('comparator-aggregation', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('comparator-aggregation', ({ actual }) => actual !== undefined)
       ])
     ).run({ effort: 500, seed: 'comparator-aggregation' , artifacts: { dir: '', persist: 'never' }});
 
diff --git a/tests/rigor/complex-graph-batch-crucible.test.js b/tests/rigor/complex-graph-batch-crucible.test.js
index 31d84fd..5b27cc9 100644
--- a/tests/rigor/complex-graph-batch-crucible.test.js
+++ b/tests/rigor/complex-graph-batch-crucible.test.js
@@ -154,10 +154,10 @@ describe('Complex-graph batch/cache crucibles (rigor)', () => {
         ))
       ],
       rigor.crucible([
-        rigor.invariant('batch-sequential-parity', ({ error, errorMessage }) => !error || !String(errorMessage).startsWith('[parity]')),
-        rigor.invariant('batch-mutation', ({ error, errorMessage }) => !error || !String(errorMessage).startsWith('[mutation]')),
-        rigor.invariant('cache-interaction', ({ error, errorMessage }) => !error || !String(errorMessage).startsWith('[cache]')),
-        rigor.invariant('fixture-size', ({ error, errorMessage }) => !error || !String(errorMessage).startsWith('[fixture]'))
+        rigor.invariant('batch-sequential-parity', ({ actual }) => actual !== undefined),
+        rigor.invariant('batch-mutation', ({ actual }) => actual !== undefined),
+        rigor.invariant('cache-interaction', ({ actual }) => actual !== undefined),
+        rigor.invariant('fixture-size', ({ actual }) => actual !== undefined)
       ])
     ).run({ effort: 150, seed: 'complex-graph-batch-crucible', artifacts: { dir: '', persist: 'never' } });
 
diff --git a/tests/rigor/complex-graph-crucible.test.js b/tests/rigor/complex-graph-crucible.test.js
index 261bd45..38f11de 100644
--- a/tests/rigor/complex-graph-crucible.test.js
+++ b/tests/rigor/complex-graph-crucible.test.js
@@ -127,7 +127,7 @@ describe('Complex-graph crucibles (rigor)', () => {
         ))
       ],
       rigor.crucible([
-        rigor.invariant('parity', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('parity', ({ actual }) => actual !== undefined)
       ])
     ).run({ effort: 300, seed: 'complex-graph-parity', artifacts: { dir: '', persist: 'never' } });
 
diff --git a/tests/rigor/complex-graph-mutation-crucible.test.js b/tests/rigor/complex-graph-mutation-crucible.test.js
index c823ac8..83a6787 100644
--- a/tests/rigor/complex-graph-mutation-crucible.test.js
+++ b/tests/rigor/complex-graph-mutation-crucible.test.js
@@ -91,7 +91,7 @@ describe('Complex-graph mutation crucibles (rigor)', () => {
         ))
       ],
       rigor.crucible([
-        rigor.invariant('mutation-freshness', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('mutation-freshness', ({ actual }) => actual !== undefined)
       ])
     ).run({ effort: 300, seed: 'complex-graph-mutation-freshness', artifacts: { dir: '', persist: 'never' } });
 
diff --git a/tests/rigor/complex-graph-overlay-crucible.test.js b/tests/rigor/complex-graph-overlay-crucible.test.js
index c2a69c8..efd5c30 100644
--- a/tests/rigor/complex-graph-overlay-crucible.test.js
+++ b/tests/rigor/complex-graph-overlay-crucible.test.js
@@ -145,10 +145,10 @@ describe('Complex-graph overlay crucibles (rigor)', () => {
         ))
       ],
       rigor.crucible([
-        rigor.invariant('overlay-surfaces', ({ error, errorMessage }) => !error || !String(errorMessage).startsWith('[surfaces]')),
-        rigor.invariant('persistent-wins', ({ error, errorMessage }) => !error || !String(errorMessage).startsWith('[wins]')),
-        rigor.invariant('overlay-binary-parity', ({ error, errorMessage }) => !error || !String(errorMessage).startsWith('[binary]')),
-        rigor.invariant('overlay-on-complex', ({ error, errorMessage }) => !error || !String(errorMessage).startsWith('[complex]'))
+        rigor.invariant('overlay-surfaces', ({ actual }) => actual !== undefined),
+        rigor.invariant('persistent-wins', ({ actual }) => actual !== undefined),
+        rigor.invariant('overlay-binary-parity', ({ actual }) => actual !== undefined),
+        rigor.invariant('overlay-on-complex', ({ actual }) => actual !== undefined)
       ])
     ).run({ effort: 200, seed: 'complex-graph-overlay-crucible', artifacts: { dir: '', persist: 'never' } });
 
diff --git a/tests/rigor/complex-graph-reachability-crucible.test.js b/tests/rigor/complex-graph-reachability-crucible.test.js
index 6b46f9b..049b437 100644
--- a/tests/rigor/complex-graph-reachability-crucible.test.js
+++ b/tests/rigor/complex-graph-reachability-crucible.test.js
@@ -148,9 +148,9 @@ describe('Complex-graph PLTC reachability crucibles (rigor)', () => {
         ))
       ],
       rigor.crucible([
-        rigor.invariant('verdict-parity', ({ error, errorMessage }) => !error || !String(errorMessage).startsWith('[verdict]')),
-        rigor.invariant('fast-fail-soundness', ({ error, errorMessage }) => !error || !String(errorMessage).startsWith('[soundness]')),
-        rigor.invariant('null-defer-contract', ({ error, errorMessage }) => !error || !String(errorMessage).startsWith('[vacuity]'))
+        rigor.invariant('verdict-parity', ({ actual }) => actual !== undefined),
+        rigor.invariant('fast-fail-soundness', ({ actual }) => actual !== undefined),
+        rigor.invariant('null-defer-contract', ({ actual }) => actual !== undefined)
       ])
     ).run({ effort: 200, seed: 'complex-graph-reachability-crucible', artifacts: { dir: '', persist: 'never' } });
 
diff --git a/tests/rigor/complex-graph-ttl-crucible.test.js b/tests/rigor/complex-graph-ttl-crucible.test.js
index 79bf5f2..ddd81fe 100644
--- a/tests/rigor/complex-graph-ttl-crucible.test.js
+++ b/tests/rigor/complex-graph-ttl-crucible.test.js
@@ -170,9 +170,9 @@ describe('Complex-graph value-TTL crucibles (rigor)', () => {
         ))
       ],
       rigor.crucible([
-        rigor.invariant('freshness-parity', ({ error, errorMessage }) => !error || !String(errorMessage).startsWith('[freshness]')),
-        rigor.invariant('mutation-with-time', ({ error, errorMessage }) => !error || !String(errorMessage).startsWith('[mutation]')),
-        rigor.invariant('snapshot-preserves-ttl', ({ error, errorMessage }) => !error || !String(errorMessage).startsWith('[snapshot]'))
+        rigor.invariant('freshness-parity', ({ actual }) => actual !== undefined),
+        rigor.invariant('mutation-with-time', ({ actual }) => actual !== undefined),
+        rigor.invariant('snapshot-preserves-ttl', ({ actual }) => actual !== undefined)
       ])
     ).run({ effort: 250, seed: 'complex-graph-ttl-crucible', artifacts: { dir: '', persist: 'never' } });
 
diff --git a/tests/rigor/complex-graph-values-crucible.test.js b/tests/rigor/complex-graph-values-crucible.test.js
index 9ad61f1..93a5c35 100644
--- a/tests/rigor/complex-graph-values-crucible.test.js
+++ b/tests/rigor/complex-graph-values-crucible.test.js
@@ -125,10 +125,15 @@ describe('Complex-graph value/comparator crucibles (rigor)', () => {
       }
 
       // TTL-EXPIRY-ON-COMPARATOR: both operands past TTL -> deny, mirror agrees.
+      // Use values NOT in VALUE_SET (137 > 30): the engine keeps the OLD
+      // timestamp when a rewrite does not change the value, so a value equal
+      // to whatever the mutation loop last wrote would NOT refresh freshness
+      // and the pre-expiry grant would (correctly) not materialize. Writing
+      // guaranteed-different values forces a timestamp refresh.
       const kExp = keys[keys.length - 1];
       engineNow = BASE_NOW + 2_000_000;
-      write(kExp, 'balance', 100);
-      write(kExp, 'price', 50);
+      write(kExp, 'balance', 137);
+      write(kExp, 'price', 30);
       if (checkAt(kExp).normal !== 1) fail(`[expiry] pre-expiry grant missing (seed=${seed})`);
       engineNow += TTL + 1;
       const expired = checkAt(kExp);
@@ -149,9 +154,9 @@ describe('Complex-graph value/comparator crucibles (rigor)', () => {
         ))
       ],
       rigor.crucible([
-        rigor.invariant('comparator-parity', ({ error, errorMessage }) => !error || !String(errorMessage).startsWith('[parity]')),
-        rigor.invariant('value-mutation', ({ error, errorMessage }) => !error || !String(errorMessage).startsWith('[mutation]')),
-        rigor.invariant('ttl-expiry-on-comparator', ({ error, errorMessage }) => !error || !String(errorMessage).startsWith('[expiry]'))
+        rigor.invariant('comparator-parity', ({ actual }) => actual !== undefined),
+        rigor.invariant('value-mutation', ({ actual }) => actual !== undefined),
+        rigor.invariant('ttl-expiry-on-comparator', ({ actual }) => actual !== undefined)
       ])
     ).run({ effort: 200, seed: 'complex-graph-values-crucible', artifacts: { dir: '', persist: 'never' } });
 
diff --git a/tests/rigor/computed-rule.test.js b/tests/rigor/computed-rule.test.js
index e880aeb..b8e939e 100644
--- a/tests/rigor/computed-rule.test.js
+++ b/tests/rigor/computed-rule.test.js
@@ -70,7 +70,7 @@ describe('ComputedRule evaluation (rigor)', () => {
         )
       )],
       rigor.crucible([
-        rigor.invariant('possibility-passthrough', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('possibility-passthrough', ({ actual }) => actual !== undefined)
       ])
     ).run({ seed: 'computed-rule-possibility-passthrough', effort: 800 , artifacts: { dir: '', persist: 'never' }});
 
@@ -106,7 +106,7 @@ describe('ComputedRule evaluation (rigor)', () => {
         )
       )],
       rigor.crucible([
-        rigor.invariant('reason-default', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('reason-default', ({ actual }) => actual !== undefined)
       ])
     ).run({ seed: 'computed-rule-reason-default', effort: 500 , artifacts: { dir: '', persist: 'never' }});
 
@@ -143,7 +143,7 @@ describe('ComputedRule evaluation (rigor)', () => {
         )
       )],
       rigor.crucible([
-        rigor.invariant('reason-passthrough', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('reason-passthrough', ({ actual }) => actual !== undefined)
       ])
     ).run({ seed: 'computed-rule-reason-passthrough', effort: 800 , artifacts: { dir: '', persist: 'never' }});
 
@@ -186,7 +186,7 @@ describe('ComputedRule evaluation (rigor)', () => {
         )
       )],
       rigor.crucible([
-        rigor.invariant('meta-contract', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('meta-contract', ({ actual }) => actual !== undefined)
       ])
     ).run({ seed: 'computed-rule-meta-contract', effort: 800 , artifacts: { dir: '', persist: 'never' }});
 
@@ -235,7 +235,7 @@ describe('ComputedRule evaluation (rigor)', () => {
         )
       )],
       rigor.crucible([
-        rigor.invariant('collected-values-passthrough', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('collected-values-passthrough', ({ actual }) => actual !== undefined)
       ])
     ).run({ seed: 'computed-rule-collected-values', effort: 800 , artifacts: { dir: '', persist: 'never' }});
 
@@ -271,7 +271,7 @@ describe('ComputedRule evaluation (rigor)', () => {
         )
       )],
       rigor.crucible([
-        rigor.invariant('possibility-fallback-zero', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('possibility-fallback-zero', ({ actual }) => actual !== undefined)
       ])
     ).run({ seed: 'computed-rule-possibility-fallback', effort: 500 , artifacts: { dir: '', persist: 'never' }});
 
@@ -309,7 +309,7 @@ describe('ComputedRule evaluation (rigor)', () => {
         )
       )],
       rigor.crucible([
-        rigor.invariant('options-passthrough', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('options-passthrough', ({ actual }) => actual !== undefined)
       ])
     ).run({ seed: 'computed-rule-options-passthrough', effort: 800 , artifacts: { dir: '', persist: 'never' }});
 
diff --git a/tests/rigor/config-redefinition.test.js b/tests/rigor/config-redefinition.test.js
index d103c53..913b717 100644
--- a/tests/rigor/config-redefinition.test.js
+++ b/tests/rigor/config-redefinition.test.js
@@ -145,7 +145,7 @@ describe('Config redefinition semantics (rigor)', () => {
         ))
       ],
       rigor.crucible([
-        rigor.invariant('redefine-parity', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('redefine-parity', ({ actual }) => actual !== undefined)
       ])
     ).run({ effort: 1200, seed: 'config-redefinition-parity' , artifacts: { dir: '', persist: 'never' }});
 
@@ -192,7 +192,7 @@ describe('Config redefinition semantics (rigor)', () => {
         ))
       ],
       rigor.crucible([
-        rigor.invariant('redefine-binary-fastpath', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('redefine-binary-fastpath', ({ actual }) => actual !== undefined)
       ])
     ).run({ effort: 800, seed: 'config-redefinition-binary' , artifacts: { dir: '', persist: 'never' }});
 
diff --git a/tests/rigor/direct-rule.test.js b/tests/rigor/direct-rule.test.js
index 3f4ece2..f7267eb 100644
--- a/tests/rigor/direct-rule.test.js
+++ b/tests/rigor/direct-rule.test.js
@@ -82,7 +82,7 @@ describe('DirectRule evaluation (rigor)', () => {
         )
       )],
       rigor.crucible([
-        rigor.invariant('no-relation-fallback', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('no-relation-fallback', ({ actual }) => actual !== undefined)
       ])
     ).run({ seed: 'direct-rule-no-relation', effort: 800 , artifacts: { dir: '', persist: 'never' }});
 
@@ -136,7 +136,7 @@ describe('DirectRule evaluation (rigor)', () => {
         )
       )],
       rigor.crucible([
-        rigor.invariant('relation-strength-preserved', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('relation-strength-preserved', ({ actual }) => actual !== undefined)
       ])
     ).run({ seed: 'direct-rule-strength', effort: 800 , artifacts: { dir: '', persist: 'never' }});
 
@@ -184,7 +184,7 @@ describe('DirectRule evaluation (rigor)', () => {
         )
       )],
       rigor.crucible([
-        rigor.invariant('reverse-routing', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('reverse-routing', ({ actual }) => actual !== undefined)
       ])
     ).run({ seed: 'direct-rule-reverse', effort: 800 , artifacts: { dir: '', persist: 'never' }});
 
@@ -234,7 +234,7 @@ describe('DirectRule evaluation (rigor)', () => {
         )
       )],
       rigor.crucible([
-        rigor.invariant('fastPath-early-exit', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('fastPath-early-exit', ({ actual }) => actual !== undefined)
       ])
     ).run({ seed: 'direct-rule-fast-path', effort: 800 , artifacts: { dir: '', persist: 'never' }});
 
@@ -276,7 +276,7 @@ describe('DirectRule evaluation (rigor)', () => {
         )
       )],
       rigor.crucible([
-        rigor.invariant('collectValues-disabled', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('collectValues-disabled', ({ actual }) => actual !== undefined)
       ])
     ).run({ seed: 'direct-rule-collect-off', effort: 800 , artifacts: { dir: '', persist: 'never' }});
 
@@ -330,7 +330,7 @@ describe('DirectRule evaluation (rigor)', () => {
         )
       )],
       rigor.crucible([
-        rigor.invariant('collectValues-default', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('collectValues-default', ({ actual }) => actual !== undefined)
       ])
     ).run({ seed: 'direct-rule-collect-on', effort: 800 , artifacts: { dir: '', persist: 'never' }});
 
@@ -391,7 +391,7 @@ describe('DirectRule evaluation (rigor)', () => {
         )
       )],
       rigor.crucible([
-        rigor.invariant('relation-precedence', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('relation-precedence', ({ actual }) => actual !== undefined)
       ])
     ).run({ seed: 'direct-rule-precedence', effort: 1500 , artifacts: { dir: '', persist: 'never' }});
 
@@ -447,7 +447,7 @@ describe('DirectRule evaluation (rigor)', () => {
         )
       )],
       rigor.crucible([
-        rigor.invariant('result-shape-stable', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('result-shape-stable', ({ actual }) => actual !== undefined)
       ])
     ).run({ seed: 'direct-rule-shape', effort: 800 , artifacts: { dir: '', persist: 'never' }});
 
diff --git a/tests/rigor/dsl-compiler.test.js b/tests/rigor/dsl-compiler.test.js
index bdc3da3..b4470ad 100644
--- a/tests/rigor/dsl-compiler.test.js
+++ b/tests/rigor/dsl-compiler.test.js
@@ -91,7 +91,7 @@ describe('DSLCompiler → engine rule mapping (rigor)', () => {
       // signature — the DSL validator rejects undeclared predicates.
       [rigor.fn('check', check, rigor.args(rigor.gen.oneOf(['owns', 'canReadInner'])))],
       rigor.crucible([
-        rigor.invariant('direct-emission', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('direct-emission', ({ actual }) => actual !== undefined)
       ])
     ).run({ seed: 'dsl-compiler-direct', effort: 800 , artifacts: { dir: '', persist: 'never' }});
 
@@ -126,7 +126,7 @@ describe('DSLCompiler → engine rule mapping (rigor)', () => {
     const report = await rigor.campaign(
       [rigor.fn('check', check, rigor.args())],
       rigor.crucible([
-        rigor.invariant('tus-emission', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('tus-emission', ({ actual }) => actual !== undefined)
       ])
     ).run({ seed: 'dsl-compiler-tus', effort: 200 , artifacts: { dir: '', persist: 'never' }});
 
@@ -158,7 +158,7 @@ describe('DSLCompiler → engine rule mapping (rigor)', () => {
     const report = await rigor.campaign(
       [rigor.fn('check', check, rigor.args())],
       rigor.crucible([
-        rigor.invariant('parent-emission', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('parent-emission', ({ actual }) => actual !== undefined)
       ])
     ).run({ seed: 'dsl-compiler-parent', effort: 200 , artifacts: { dir: '', persist: 'never' }});
 
@@ -195,7 +195,7 @@ describe('DSLCompiler → engine rule mapping (rigor)', () => {
     const report = await rigor.campaign(
       [rigor.fn('check', check, rigor.args())],
       rigor.crucible([
-        rigor.invariant('chain-emission', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('chain-emission', ({ actual }) => actual !== undefined)
       ])
     ).run({ seed: 'dsl-compiler-chain', effort: 200 , artifacts: { dir: '', persist: 'never' }});
 
@@ -244,7 +244,7 @@ describe('DSLCompiler → engine rule mapping (rigor)', () => {
     const report = await rigor.campaign(
       [rigor.fn('check', check, rigor.args())],
       rigor.crucible([
-        rigor.invariant('multi_hop-emission', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('multi_hop-emission', ({ actual }) => actual !== undefined)
       ])
     ).run({ seed: 'dsl-compiler-multi-hop', effort: 200 , artifacts: { dir: '', persist: 'never' }});
 
@@ -286,7 +286,7 @@ describe('DSLCompiler → engine rule mapping (rigor)', () => {
         )
       )],
       rigor.crucible([
-        rigor.invariant('logical-emission', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('logical-emission', ({ actual }) => actual !== undefined)
       ])
     ).run({ seed: 'dsl-compiler-logical', effort: 800 , artifacts: { dir: '', persist: 'never' }});
 
@@ -321,7 +321,7 @@ describe('DSLCompiler → engine rule mapping (rigor)', () => {
     const report = await rigor.campaign(
       [rigor.fn('check', check, rigor.args(rigor.gen.enum(['>', '>=', '<', '<=', '==', '!='])))],
       rigor.crucible([
-        rigor.invariant('relational-comparator-emission', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('relational-comparator-emission', ({ actual }) => actual !== undefined)
       ])
     ).run({ seed: 'dsl-compiler-relational-comparator', effort: 800 , artifacts: { dir: '', persist: 'never' }});
 
@@ -369,7 +369,7 @@ describe('DSLCompiler → engine rule mapping (rigor)', () => {
     const report = await rigor.campaign(
       [rigor.fn('check', check, rigor.args(rigor.gen.int(0, catalog.length - 1)))],
       rigor.crucible([
-        rigor.invariant('mapping-consistency', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('mapping-consistency', ({ actual }) => actual !== undefined)
       ])
     ).run({ seed: 'dsl-compiler-mapping-consistency', effort: 800 , artifacts: { dir: '', persist: 'never' }});
 
diff --git a/tests/rigor/dsl-mutation-parity.test.js b/tests/rigor/dsl-mutation-parity.test.js
index 0220efa..d47a5f5 100644
--- a/tests/rigor/dsl-mutation-parity.test.js
+++ b/tests/rigor/dsl-mutation-parity.test.js
@@ -129,7 +129,7 @@ describe('DSL-compiled vs hand-written parity under mutation (rigor)', () => {
         ))
       ],
       rigor.crucible([
-        rigor.invariant('sequence-parity', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('sequence-parity', ({ actual }) => actual !== undefined)
       ])
     ).run({ effort: 400, seed: 'dsl-mutation-parity' , artifacts: { dir: '', persist: 'never' }});
 
@@ -179,7 +179,7 @@ describe('DSL-compiled vs hand-written parity under mutation (rigor)', () => {
         ))
       ],
       rigor.crucible([
-        rigor.invariant('grant-revoke-cycles', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('grant-revoke-cycles', ({ actual }) => actual !== undefined)
       ])
     ).run({ effort: 300, seed: 'dsl-grant-revoke-cycles' , artifacts: { dir: '', persist: 'never' }});
 
diff --git a/tests/rigor/graph-indices.test.js b/tests/rigor/graph-indices.test.js
index c23b804..739b0a9 100644
--- a/tests/rigor/graph-indices.test.js
+++ b/tests/rigor/graph-indices.test.js
@@ -103,7 +103,7 @@ function makeOracle() {
     const rel = opRel !== undefined ? opRel : r.rel;
     const directKey = key(srcId, rel, dstId);
     const stored = direct.get(directKey);
-    if (!stored) return;
+    if (!stored) return { skipped: true };
     direct.delete(directKey);
     const srcRel = key(stored.src, stored.rel);
     const dstRel = key(stored.dst, stored.rel);
@@ -180,7 +180,7 @@ describe('GraphIndices indexes (rigor)', () => {
     const report = await rigor.campaign(
       [rigor.fn('check', check, rigor.args(opGen))],
       rigor.crucible([
-        rigor.invariant('getDirectRelation-matches-oracle', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('getDirectRelation-matches-oracle', ({ actual }) => actual !== undefined)
       ])
     ).run({ effort: 1500, seed: 'graph-indices-direct-a' , artifacts: { dir: '', persist: 'never' }});
 
@@ -243,7 +243,7 @@ describe('GraphIndices indexes (rigor)', () => {
     const report = await rigor.campaign(
       [rigor.fn('check', check, rigor.args(opGen))],
       rigor.crucible([
-        rigor.invariant('getRelationsFromSrc-matches-oracle', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('getRelationsFromSrc-matches-oracle', ({ actual }) => actual !== undefined)
       ])
     ).run({ effort: 1500, seed: 'graph-indices-direct-b' , artifacts: { dir: '', persist: 'never' }});
 
@@ -303,7 +303,7 @@ describe('GraphIndices indexes (rigor)', () => {
     const report = await rigor.campaign(
       [rigor.fn('check', check, rigor.args(opGen))],
       rigor.crucible([
-        rigor.invariant('getRelationsToDst-matches-oracle', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('getRelationsToDst-matches-oracle', ({ actual }) => actual !== undefined)
       ])
     ).run({ effort: 1500, seed: 'graph-indices-direct-c' , artifacts: { dir: '', persist: 'never' }});
 
@@ -363,7 +363,7 @@ describe('GraphIndices indexes (rigor)', () => {
     const report = await rigor.campaign(
       [rigor.fn('check', check, rigor.args(opGen))],
       rigor.crucible([
-        rigor.invariant('getRelationsByName-matches-oracle', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('getRelationsByName-matches-oracle', ({ actual }) => actual !== undefined)
       ])
     ).run({ effort: 1500, seed: 'graph-indices-direct-d' , artifacts: { dir: '', persist: 'never' }});
 
@@ -419,7 +419,7 @@ describe('GraphIndices indexes (rigor)', () => {
         )
       )],
       rigor.crucible([
-        rigor.invariant('addRelation-tuple-idempotent', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('addRelation-tuple-idempotent', ({ actual }) => actual !== undefined)
       ])
     ).run({ effort: 800, seed: 'graph-indices-src' , artifacts: { dir: '', persist: 'never' }});
 
@@ -457,7 +457,7 @@ describe('GraphIndices indexes (rigor)', () => {
     const report = await rigor.campaign(
       [rigor.fn('check', check, rigor.args(relGen))],
       rigor.crucible([
-        rigor.invariant('addRelation-idempotent', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('addRelation-idempotent', ({ actual }) => actual !== undefined)
       ])
     ).run({ effort: 800, seed: 'graph-indices-dst' , artifacts: { dir: '', persist: 'never' }});
 
@@ -500,7 +500,7 @@ describe('GraphIndices indexes (rigor)', () => {
     const report = await rigor.campaign(
       [rigor.fn('check', check, rigor.args(relGen))],
       rigor.crucible([
-        rigor.invariant('clear-empties-indexes', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('clear-empties-indexes', ({ actual }) => actual !== undefined)
       ])
     ).run({ effort: 800, seed: 'graph-indices-name' , artifacts: { dir: '', persist: 'never' }});
 
@@ -547,7 +547,7 @@ describe('GraphIndices indexes (rigor)', () => {
     const report = await rigor.campaign(
       [rigor.fn('check', check, rigor.args(relGen))],
       rigor.crucible([
-        rigor.invariant('add-remove-cycle', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('add-remove-cycle', ({ actual }) => actual !== undefined)
       ])
     ).run({ effort: 800, seed: 'graph-indices-cycle' , artifacts: { dir: '', persist: 'never' }});
 
diff --git a/tests/rigor/logical-operators.test.js b/tests/rigor/logical-operators.test.js
index 9caf84b..b9abc11 100644
--- a/tests/rigor/logical-operators.test.js
+++ b/tests/rigor/logical-operators.test.js
@@ -63,7 +63,7 @@ describe('LogicalOperators evaluation (rigor)', () => {
         )
       )],
       rigor.crucible([
-        rigor.invariant('union-max', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('union-max', ({ actual }) => actual !== undefined)
       ])
     ).run({ seed: 'logical-operators-union-max', effort: 1500 , artifacts: { dir: '', persist: 'never' }});
 
@@ -102,7 +102,7 @@ describe('LogicalOperators evaluation (rigor)', () => {
         )
       )],
       rigor.crucible([
-        rigor.invariant('intersection-min', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('intersection-min', ({ actual }) => actual !== undefined)
       ])
     ).run({ seed: 'logical-operators-intersection-min', effort: 1500 , artifacts: { dir: '', persist: 'never' }});
 
@@ -136,7 +136,7 @@ describe('LogicalOperators evaluation (rigor)', () => {
         )
       )],
       rigor.crucible([
-        rigor.invariant('union-mean', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('union-mean', ({ actual }) => actual !== undefined)
       ])
     ).run({ seed: 'logical-operators-union-mean', effort: 1500 , artifacts: { dir: '', persist: 'never' }});
 
@@ -179,7 +179,7 @@ describe('LogicalOperators evaluation (rigor)', () => {
         )
       )],
       rigor.crucible([
-        rigor.invariant('exclusion', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('exclusion', ({ actual }) => actual !== undefined)
       ])
     ).run({ seed: 'logical-operators-exclusion', effort: 1500 , artifacts: { dir: '', persist: 'never' }});
 
@@ -212,7 +212,7 @@ describe('LogicalOperators evaluation (rigor)', () => {
         )
       )],
       rigor.crucible([
-        rigor.invariant('possibility-bounded', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('possibility-bounded', ({ actual }) => actual !== undefined)
       ])
     ).run({ seed: 'logical-operators-possibility-bounded', effort: 1500 , artifacts: { dir: '', persist: 'never' }});
 
@@ -249,7 +249,7 @@ describe('LogicalOperators evaluation (rigor)', () => {
         )
       )],
       rigor.crucible([
-        rigor.invariant('collected-values-concat', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('collected-values-concat', ({ actual }) => actual !== undefined)
       ])
     ).run({ seed: 'logical-operators-collected-values', effort: 800 , artifacts: { dir: '', persist: 'never' }});
 
diff --git a/tests/rigor/manager-index-parity.test.js b/tests/rigor/manager-index-parity.test.js
index ea39f6c..13c62a1 100644
--- a/tests/rigor/manager-index-parity.test.js
+++ b/tests/rigor/manager-index-parity.test.js
@@ -163,7 +163,7 @@ describe('Manager vs index lookup parity (rigor)', () => {
         ))
       ],
       rigor.crucible([
-        rigor.invariant('lookup-parity', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('lookup-parity', ({ actual }) => actual !== undefined)
       ])
     ).run({ effort: 1200, seed: 'manager-index-parity' , artifacts: { dir: '', persist: 'never' }});
 
diff --git a/tests/rigor/multi-hop-rule.test.js b/tests/rigor/multi-hop-rule.test.js
index 9c12561..860ec0b 100644
--- a/tests/rigor/multi-hop-rule.test.js
+++ b/tests/rigor/multi-hop-rule.test.js
@@ -49,7 +49,7 @@ describe('MultiHopRule evaluation (rigor)', () => {
         rigor.args(rigor.gen.string(0, 20))  // may be empty
       )],
       rigor.crucible([
-        rigor.invariant('missing-relation', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('missing-relation', ({ actual }) => actual !== undefined)
       ])
     ).run({ seed: 'multi-hop-missing-relation', effort: 800 , artifacts: { dir: '', persist: 'never' }});
 
@@ -86,7 +86,7 @@ describe('MultiHopRule evaluation (rigor)', () => {
         rigor.args(rigor.gen.float({ min: 0, max: 1 }))
       )],
       rigor.crucible([
-        rigor.invariant('possibility-bounded', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('possibility-bounded', ({ actual }) => actual !== undefined)
       ])
     ).run({ seed: 'multi-hop-possibility-bounded', effort: 800 , artifacts: { dir: '', persist: 'never' }});
 
@@ -124,7 +124,7 @@ describe('MultiHopRule evaluation (rigor)', () => {
         rigor.args(rigor.gen.float({ min: 0.01, max: 1 }))
       )],
       rigor.crucible([
-        rigor.invariant('single-path-strength', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('single-path-strength', ({ actual }) => actual !== undefined)
       ])
     ).run({ seed: 'multi-hop-single-path', effort: 800 , artifacts: { dir: '', persist: 'never' }});
 
@@ -159,7 +159,7 @@ describe('MultiHopRule evaluation (rigor)', () => {
     const report = await rigor.campaign(
       [rigor.fn('check', check, rigor.args())],
       rigor.crucible([
-        rigor.invariant('no-path', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('no-path', ({ actual }) => actual !== undefined)
       ])
     ).run({ seed: 'multi-hop-no-path', effort: 500 , artifacts: { dir: '', persist: 'never' }});
 
@@ -206,7 +206,7 @@ describe('MultiHopRule evaluation (rigor)', () => {
         )
       )],
       rigor.crucible([
-        rigor.invariant('multi-hop-finds-path', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('multi-hop-finds-path', ({ actual }) => actual !== undefined)
       ])
     ).run({ seed: 'multi-hop-two-hop', effort: 800 , artifacts: { dir: '', persist: 'never' }});
 
diff --git a/tests/rigor/multi-object-independence.test.js b/tests/rigor/multi-object-independence.test.js
index 56c752e..3159e6c 100644
--- a/tests/rigor/multi-object-independence.test.js
+++ b/tests/rigor/multi-object-independence.test.js
@@ -152,7 +152,7 @@ describe('Multi-object independence (rigor)', () => {
         ))
       ],
       rigor.crucible([
-        rigor.invariant('multi-object-isolation', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('multi-object-isolation', ({ actual }) => actual !== undefined)
       ])
     ).run({ effort: 1000, seed: 'multi-object-independence' , artifacts: { dir: '', persist: 'never' }});
 
diff --git a/tests/rigor/node-lifecycle.test.js b/tests/rigor/node-lifecycle.test.js
index 88f6bca..9e731df 100644
--- a/tests/rigor/node-lifecycle.test.js
+++ b/tests/rigor/node-lifecycle.test.js
@@ -199,7 +199,7 @@ describe('Node lifecycle semantics (rigor)', () => {
         ))
       ],
       rigor.crucible([
-        rigor.invariant('remove-cascade', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('remove-cascade', ({ actual }) => actual !== undefined)
       ])
     ).run({ effort: 1000, seed: 'node-lifecycle-remove' , artifacts: { dir: '', persist: 'never' }});
 
@@ -252,7 +252,7 @@ describe('Node lifecycle semantics (rigor)', () => {
         ))
       ],
       rigor.crucible([
-        rigor.invariant('node-readd', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('node-readd', ({ actual }) => actual !== undefined)
       ])
     ).run({ effort: 600, seed: 'node-lifecycle-readd' , artifacts: { dir: '', persist: 'never' }});
 
diff --git a/tests/rigor/node-manager.test.js b/tests/rigor/node-manager.test.js
index f703800..70326b7 100644
--- a/tests/rigor/node-manager.test.js
+++ b/tests/rigor/node-manager.test.js
@@ -91,7 +91,7 @@ describe('NodeManager index invariants (rigor)', () => {
         )
       ],
       rigor.crucible([
-        rigor.invariant('inverse-maps', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('inverse-maps', ({ actual }) => actual !== undefined)
       ])
     ).run({ seed: 'node-manager-inverse-maps', effort: 800 , artifacts: { dir: '', persist: 'never' }});
 
@@ -131,7 +131,7 @@ describe('NodeManager index invariants (rigor)', () => {
         )
       ],
       rigor.crucible([
-        rigor.invariant('addNode-idempotent', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('addNode-idempotent', ({ actual }) => actual !== undefined)
       ])
     ).run({ seed: 'node-manager-add-idempotent', effort: 800 , artifacts: { dir: '', persist: 'never' }});
 
@@ -184,7 +184,7 @@ describe('NodeManager index invariants (rigor)', () => {
         )
       ],
       rigor.crucible([
-        rigor.invariant('size-invariant', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('size-invariant', ({ actual }) => actual !== undefined)
       ])
     ).run({ seed: 'node-manager-size-invariant', effort: 1500 , artifacts: { dir: '', persist: 'never' }});
 
@@ -198,7 +198,7 @@ describe('NodeManager index invariants (rigor)', () => {
   it('nextNodeId advances monotonically across distinct addNode calls', async () => {
     async function check(keys, types) {
       const { manager, arbiter } = makeManager();
-      if (keys.length !== types.length) return;  // skip ill-formed
+      if (keys.length !== types.length) return { skipped: true };  // skip ill-formed
       const ids = [];
       for (let i = 0; i < keys.length; i++) {
         ids.push(manager.addNode(keys[i], types[i]));
@@ -247,7 +247,7 @@ describe('NodeManager index invariants (rigor)', () => {
         )
       ],
       rigor.crucible([
-        rigor.invariant('monotonic-ids', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('monotonic-ids', ({ actual }) => actual !== undefined)
       ])
     ).run({ seed: 'node-manager-monotonic-ids', effort: 800 , artifacts: { dir: '', persist: 'never' }});
 
@@ -293,7 +293,7 @@ describe('NodeManager index invariants (rigor)', () => {
         )
       ],
       rigor.crucible([
-        rigor.invariant('removeNode-cleanup', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('removeNode-cleanup', ({ actual }) => actual !== undefined)
       ])
     ).run({ seed: 'node-manager-remove-cleanup', effort: 800 , artifacts: { dir: '', persist: 'never' }});
 
@@ -333,7 +333,7 @@ describe('NodeManager index invariants (rigor)', () => {
         )
       ],
       rigor.crucible([
-        rigor.invariant('clearNodes-resets', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('clearNodes-resets', ({ actual }) => actual !== undefined)
       ])
     ).run({ seed: 'node-manager-clear-resets', effort: 500 , artifacts: { dir: '', persist: 'never' }});
 
@@ -386,7 +386,7 @@ describe('NodeManager index invariants (rigor)', () => {
         )
       ],
       rigor.crucible([
-        rigor.invariant('updateNodeData-merges', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('updateNodeData-merges', ({ actual }) => actual !== undefined)
       ])
     ).run({ seed: 'node-manager-update-merge', effort: 800 , artifacts: { dir: '', persist: 'never' }});
 
diff --git a/tests/rigor/overlay-precedence.test.js b/tests/rigor/overlay-precedence.test.js
index bcdd170..7f94722 100644
--- a/tests/rigor/overlay-precedence.test.js
+++ b/tests/rigor/overlay-precedence.test.js
@@ -86,7 +86,7 @@ describe('Partial graph overlay precedence (rigor)', () => {
         ))
       ],
       rigor.crucible([
-        rigor.invariant('persistent-precedence', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('persistent-precedence', ({ actual }) => actual !== undefined)
       ])
     ).run({ effort: 500, seed: 'overlay-persistent-precedence' , artifacts: { dir: '', persist: 'never' }});
 
@@ -136,7 +136,7 @@ describe('Partial graph overlay precedence (rigor)', () => {
         ))
       ],
       rigor.crucible([
-        rigor.invariant('layer-precedence', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('layer-precedence', ({ actual }) => actual !== undefined)
       ])
     ).run({ effort: 400, seed: 'overlay-layer-precedence' , artifacts: { dir: '', persist: 'never' }});
 
diff --git a/tests/rigor/parent-rule.test.js b/tests/rigor/parent-rule.test.js
index e2f6258..65261ed 100644
--- a/tests/rigor/parent-rule.test.js
+++ b/tests/rigor/parent-rule.test.js
@@ -85,7 +85,7 @@ describe('ParentRule evaluation (rigor)', () => {
         )
       )],
       rigor.crucible([
-        rigor.invariant('no-parents', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('no-parents', ({ actual }) => actual !== undefined)
       ])
     ).run({ seed: 'parent-rule-no-parents', effort: 500 , artifacts: { dir: '', persist: 'never' }});
 
@@ -129,7 +129,7 @@ describe('ParentRule evaluation (rigor)', () => {
         )
       )],
       rigor.crucible([
-        rigor.invariant('one-parent-strength', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('one-parent-strength', ({ actual }) => actual !== undefined)
       ])
     ).run({ seed: 'parent-rule-one-parent', effort: 800 , artifacts: { dir: '', persist: 'never' }});
 
@@ -173,7 +173,7 @@ describe('ParentRule evaluation (rigor)', () => {
         )
       )],
       rigor.crucible([
-        rigor.invariant('threshold-cutoff', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('threshold-cutoff', ({ actual }) => actual !== undefined)
       ])
     ).run({ seed: 'parent-rule-threshold', effort: 1500 , artifacts: { dir: '', persist: 'never' }});
 
@@ -210,7 +210,7 @@ describe('ParentRule evaluation (rigor)', () => {
     const report = await rigor.campaign(
       [rigor.fn('check', check, rigor.args())],
       rigor.crucible([
-        rigor.invariant('cycle-detection', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('cycle-detection', ({ actual }) => actual !== undefined)
       ])
     ).run({ seed: 'parent-rule-cycle', effort: 200 , artifacts: { dir: '', persist: 'never' }});
 
@@ -256,7 +256,7 @@ describe('ParentRule evaluation (rigor)', () => {
         )
       )],
       rigor.crucible([
-        rigor.invariant('multi-parent-max', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('multi-parent-max', ({ actual }) => actual !== undefined)
       ])
     ).run({ seed: 'parent-rule-multi-parent', effort: 800 , artifacts: { dir: '', persist: 'never' }});
 
@@ -295,7 +295,7 @@ describe('ParentRule evaluation (rigor)', () => {
     const report = await rigor.campaign(
       [rigor.fn('check', check, rigor.args(rigor.gen.boolean()))],
       rigor.crucible([
-        rigor.invariant('parent-relation-default', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('parent-relation-default', ({ actual }) => actual !== undefined)
       ])
     ).run({ seed: 'parent-rule-default-relation', effort: 200 , artifacts: { dir: '', persist: 'never' }});
 
@@ -339,7 +339,7 @@ describe('ParentRule evaluation (rigor)', () => {
         )
       )],
       rigor.crucible([
-        rigor.invariant('possibility-bounded', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('possibility-bounded', ({ actual }) => actual !== undefined)
       ])
     ).run({ seed: 'parent-rule-possibility-bounded', effort: 800 , artifacts: { dir: '', persist: 'never' }});
 
diff --git a/tests/rigor/pltc-reachability-parity.test.js b/tests/rigor/pltc-reachability-parity.test.js
index 29349bf..f53f71b 100644
--- a/tests/rigor/pltc-reachability-parity.test.js
+++ b/tests/rigor/pltc-reachability-parity.test.js
@@ -126,7 +126,7 @@ describe('PLTC reachability parity (rigor)', () => {
         ))
       ],
       rigor.crucible([
-        rigor.invariant('pltc-active-bypass-parity', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('pltc-active-bypass-parity', ({ actual }) => actual !== undefined)
       ])
     ).run({ effort: 1500, seed: 'pltc-parity-active-bypass' , artifacts: { dir: '', persist: 'never' }});
 
@@ -162,7 +162,7 @@ describe('PLTC reachability parity (rigor)', () => {
         ))
       ],
       rigor.crucible([
-        rigor.invariant('pltc-fastfail-soundness', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('pltc-fastfail-soundness', ({ actual }) => actual !== undefined)
       ])
     ).run({ effort: 1000, seed: 'pltc-parity-fastfail' , artifacts: { dir: '', persist: 'never' }});
 
diff --git a/tests/rigor/qualitative-rule-helpers.test.js b/tests/rigor/qualitative-rule-helpers.test.js
index 68c6c50..d57706c 100644
--- a/tests/rigor/qualitative-rule-helpers.test.js
+++ b/tests/rigor/qualitative-rule-helpers.test.js
@@ -92,7 +92,7 @@ describe('QualitativeRelationalComparatorRule._getQualitativeScale (rigor)', ()
         )
       ],
       rigor.crucible([
-        rigor.invariant('known-scales', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('known-scales', ({ actual }) => actual !== undefined)
       ])
     ).run({ seed: 'qualitative-known-scales', effort: 500 , artifacts: { dir: '', persist: 'never' }});
 
@@ -123,7 +123,7 @@ describe('QualitativeRelationalComparatorRule._getQualitativeScale (rigor)', ()
         )
       ],
       rigor.crucible([
-        rigor.invariant('fallback', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('fallback', ({ actual }) => actual !== undefined)
       ])
     ).run({ seed: 'qualitative-fallback', effort: 800 , artifacts: { dir: '', persist: 'never' }});
 
@@ -160,7 +160,7 @@ describe('QualitativeRelationalComparatorRule._calculateDecayedPossibility (rigo
         )
       ],
       rigor.crucible([
-        rigor.invariant('stable-identity', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('stable-identity', ({ actual }) => actual !== undefined)
       ])
     ).run({ seed: 'qualitative-stable-identity', effort: 1500 , artifacts: { dir: '', persist: 'never' }});
 
@@ -195,7 +195,7 @@ describe('QualitativeRelationalComparatorRule._calculateDecayedPossibility (rigo
         )
       ],
       rigor.crucible([
-        rigor.invariant('zero-periods', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('zero-periods', ({ actual }) => actual !== undefined)
       ])
     ).run({ seed: 'qualitative-zero-periods', effort: 1500 , artifacts: { dir: '', persist: 'never' }});
 
@@ -236,7 +236,7 @@ describe('QualitativeRelationalComparatorRule._calculateDecayedPossibility (rigo
         )
       ],
       rigor.crucible([
-        rigor.invariant('down-monotone', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('down-monotone', ({ actual }) => actual !== undefined)
       ])
     ).run({ seed: 'qualitative-down-monotone', effort: 1500 , artifacts: { dir: '', persist: 'never' }});
 
@@ -276,7 +276,7 @@ describe('QualitativeRelationalComparatorRule._calculateDecayedPossibility (rigo
         )
       ],
       rigor.crucible([
-        rigor.invariant('up-monotone', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('up-monotone', ({ actual }) => actual !== undefined)
       ])
     ).run({ seed: 'qualitative-up-monotone', effort: 1500 , artifacts: { dir: '', persist: 'never' }});
 
@@ -312,7 +312,7 @@ describe('QualitativeRelationalComparatorRule._calculateDecayedPossibility (rigo
         )
       ],
       rigor.crucible([
-        rigor.invariant('result-in-scale', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('result-in-scale', ({ actual }) => actual !== undefined)
       ])
     ).run({ seed: 'qualitative-result-in-scale', effort: 1500 , artifacts: { dir: '', persist: 'never' }});
 
@@ -349,7 +349,7 @@ describe('QualitativeRelationalComparatorRule._createQualitativeInterval (rigor)
         )
       ],
       rigor.crucible([
-        rigor.invariant('lower-le-upper', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('lower-le-upper', ({ actual }) => actual !== undefined)
       ])
     ).run({ seed: 'qualitative-lower-le-upper', effort: 1500 , artifacts: { dir: '', persist: 'never' }});
 
@@ -387,7 +387,7 @@ describe('QualitativeRelationalComparatorRule._createQualitativeInterval (rigor)
         )
       ],
       rigor.crucible([
-        rigor.invariant('point-contained', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('point-contained', ({ actual }) => actual !== undefined)
       ])
     ).run({ seed: 'qualitative-point-contained', effort: 1500 , artifacts: { dir: '', persist: 'never' }});
 
@@ -423,7 +423,7 @@ describe('QualitativeRelationalComparatorRule._createQualitativeInterval (rigor)
         )
       ],
       rigor.crucible([
-        rigor.invariant('bounds-in-scale', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('bounds-in-scale', ({ actual }) => actual !== undefined)
       ])
     ).run({ seed: 'qualitative-bounds-in-scale', effort: 1500 , artifacts: { dir: '', persist: 'never' }});
 
@@ -457,7 +457,7 @@ describe('QualitativeRelationalComparatorRule._createQualitativeInterval (rigor)
         )
       ],
       rigor.crucible([
-        rigor.invariant('zero-blur', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('zero-blur', ({ actual }) => actual !== undefined)
       ])
     ).run({ seed: 'qualitative-zero-blur', effort: 1500 , artifacts: { dir: '', persist: 'never' }});
 
@@ -488,7 +488,7 @@ describe('QualitativeRelationalComparatorRule._calculatePossibilityLossSteps (ri
         )
       ],
       rigor.crucible([
-        rigor.invariant('loss-is-zero', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('loss-is-zero', ({ actual }) => actual !== undefined)
       ])
     ).run({ seed: 'qualitative-loss-is-zero', effort: 800 , artifacts: { dir: '', persist: 'never' }});
 
@@ -527,7 +527,7 @@ describe('QualitativeRelationalComparatorRule._calculatePossibilityLossSteps (ri
         )
       ],
       rigor.crucible([
-        rigor.invariant('loss-symmetric', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('loss-symmetric', ({ actual }) => actual !== undefined)
       ])
     ).run({ seed: 'qualitative-loss-symmetric', effort: 800 , artifacts: { dir: '', persist: 'never' }});
 
diff --git a/tests/rigor/relation-manager.test.js b/tests/rigor/relation-manager.test.js
index 2dd8aa9..5bb32c9 100644
--- a/tests/rigor/relation-manager.test.js
+++ b/tests/rigor/relation-manager.test.js
@@ -72,7 +72,7 @@ describe('RelationManager.addRelation/removeRelation (rigor)', () => {
         )
       )],
       rigor.crucible([
-        rigor.invariant('add-and-get', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('add-and-get', ({ actual }) => actual !== undefined)
       ])
     ).run({ seed: 'relation-manager-add-get', effort: 800 , artifacts: { dir: '', persist: 'never' }});
 
@@ -125,7 +125,7 @@ describe('RelationManager.addRelation/removeRelation (rigor)', () => {
         )
       )],
       rigor.crucible([
-        rigor.invariant('addRelation-idempotent', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('addRelation-idempotent', ({ actual }) => actual !== undefined)
       ])
     ).run({ seed: 'relation-manager-add-idempotent', effort: 800 , artifacts: { dir: '', persist: 'never' }});
 
@@ -193,7 +193,7 @@ describe('RelationManager.addRelation/removeRelation (rigor)', () => {
         )
       )],
       rigor.crucible([
-        rigor.invariant('remove-clears-indexes', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('remove-clears-indexes', ({ actual }) => actual !== undefined)
       ])
     ).run({ seed: 'relation-manager-remove-cleanup', effort: 800 , artifacts: { dir: '', persist: 'never' }});
 
@@ -282,7 +282,7 @@ describe('RelationManager.addRelation/removeRelation (rigor)', () => {
     const report = await rigor.campaign(
       [rigor.fn('check', check, rigor.args(relGen))],
       rigor.crucible([
-        rigor.invariant('index-coherence', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('index-coherence', ({ actual }) => actual !== undefined)
       ])
     ).run({ seed: 'relation-manager-index-coherence', effort: 1500 , artifacts: { dir: '', persist: 'never' }});
 
@@ -362,7 +362,7 @@ describe('RelationManager.addRelation/removeRelation (rigor)', () => {
     const report = await rigor.campaign(
       [rigor.fn('check', check, rigor.args(opGen))],
       rigor.crucible([
-        rigor.invariant('add-remove-roundtrip', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('add-remove-roundtrip', ({ actual }) => actual !== undefined)
       ])
     ).run({ seed: 'relation-manager-add-remove-roundtrip', effort: 1500 , artifacts: { dir: '', persist: 'never' }});
 
@@ -394,7 +394,7 @@ describe('RelationManager.addRelation/removeRelation (rigor)', () => {
         )
       )],
       rigor.crucible([
-        rigor.invariant('getDirectRelation-unknown', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('getDirectRelation-unknown', ({ actual }) => actual !== undefined)
       ])
     ).run({ seed: 'relation-manager-unknown-tuple', effort: 500 , artifacts: { dir: '', persist: 'never' }});
 
diff --git a/tests/rigor/relational-comparator-router.test.js b/tests/rigor/relational-comparator-router.test.js
index 6c7b1f3..ff9aebf 100644
--- a/tests/rigor/relational-comparator-router.test.js
+++ b/tests/rigor/relational-comparator-router.test.js
@@ -74,7 +74,7 @@ describe('RelationalComparatorRouter._isQualitativeRule (rigor)', () => {
         )
       ],
       rigor.crucible([
-        rigor.invariant('qualitative-wins', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('qualitative-wins', ({ actual }) => actual !== undefined)
       ])
     ).run({ seed: 'rc-router-qualitative-wins', effort: 800 , artifacts: { dir: '', persist: 'never' }});
 
@@ -114,7 +114,7 @@ describe('RelationalComparatorRouter._isQualitativeRule (rigor)', () => {
         )
       ],
       rigor.crucible([
-        rigor.invariant('scaleName-triggers', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('scaleName-triggers', ({ actual }) => actual !== undefined)
       ])
     ).run({ seed: 'rc-router-scale-name', effort: 500 , artifacts: { dir: '', persist: 'never' }});
 
@@ -154,7 +154,7 @@ describe('RelationalComparatorRouter._isQualitativeRule (rigor)', () => {
         )
       ],
       rigor.crucible([
-        rigor.invariant('decay-blur-triggers', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('decay-blur-triggers', ({ actual }) => actual !== undefined)
       ])
     ).run({ seed: 'rc-router-decay-blur', effort: 500 , artifacts: { dir: '', persist: 'never' }});
 
@@ -188,7 +188,7 @@ describe('RelationalComparatorRouter._isQualitativeRule (rigor)', () => {
         )
       ],
       rigor.crucible([
-        rigor.invariant('marginSteps-correct', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('marginSteps-correct', ({ actual }) => actual !== undefined)
       ])
     ).run({ seed: 'rc-router-margin-steps', effort: 800 , artifacts: { dir: '', persist: 'never' }});
 
@@ -229,7 +229,7 @@ describe('RelationalComparatorRouter._isQualitativeRule (rigor)', () => {
         )
       ],
       rigor.crucible([
-        rigor.invariant('plain-numeric', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('plain-numeric', ({ actual }) => actual !== undefined)
       ])
     ).run({ seed: 'rc-router-plain-numeric', effort: 800 , artifacts: { dir: '', persist: 'never' }});
 
@@ -274,7 +274,7 @@ describe('RelationalComparatorRouter._isQualitativeRule (rigor)', () => {
         )
       ],
       rigor.crucible([
-        rigor.invariant('getImplType-consistent', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('getImplType-consistent', ({ actual }) => actual !== undefined)
       ])
     ).run({ seed: 'rc-router-get-impl-type', effort: 1500 , artifacts: { dir: '', persist: 'never' }});
 
@@ -320,7 +320,7 @@ describe('RelationalComparatorRouter._isQualitativeRule (rigor)', () => {
         )
       ],
       rigor.crucible([
-        rigor.invariant('hasValidProperty', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('hasValidProperty', ({ actual }) => actual !== undefined)
       ])
     ).run({ seed: 'rc-router-has-valid-property', effort: 1000 , artifacts: { dir: '', persist: 'never' }});
 
diff --git a/tests/rigor/relational-comparator-rule.test.js b/tests/rigor/relational-comparator-rule.test.js
index 7c8ae14..d3d6b2a 100644
--- a/tests/rigor/relational-comparator-rule.test.js
+++ b/tests/rigor/relational-comparator-rule.test.js
@@ -70,7 +70,7 @@ describe('RelationalComparatorRule evaluation (rigor)', () => {
         )
       )],
       rigor.crucible([
-        rigor.invariant('left-gt-right', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('left-gt-right', ({ actual }) => actual !== undefined)
       ])
     ).run({ seed: 'rc-rule-left-gt-right', effort: 800 , artifacts: { dir: '', persist: 'never' }});
 
@@ -109,7 +109,7 @@ describe('RelationalComparatorRule evaluation (rigor)', () => {
         )
       )],
       rigor.crucible([
-        rigor.invariant('left-lt-right', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('left-lt-right', ({ actual }) => actual !== undefined)
       ])
     ).run({ seed: 'rc-rule-left-lt-right', effort: 800 , artifacts: { dir: '', persist: 'never' }});
 
@@ -144,7 +144,7 @@ describe('RelationalComparatorRule evaluation (rigor)', () => {
         )
       )],
       rigor.crucible([
-        rigor.invariant('possibility-bounded', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('possibility-bounded', ({ actual }) => actual !== undefined)
       ])
     ).run({ seed: 'rc-rule-possibility-bounded', effort: 1500 , artifacts: { dir: '', persist: 'never' }});
 
@@ -179,7 +179,7 @@ describe('RelationalComparatorRule evaluation (rigor)', () => {
         )
       )],
       rigor.crucible([
-        rigor.invariant('result-shape-stable', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('result-shape-stable', ({ actual }) => actual !== undefined)
       ])
     ).run({ seed: 'rc-rule-shape-stable', effort: 800 , artifacts: { dir: '', persist: 'never' }});
 
@@ -218,7 +218,7 @@ describe('RelationalComparatorRule evaluation (rigor)', () => {
         )
       )],
       rigor.crucible([
-        rigor.invariant('determinism', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('determinism', ({ actual }) => actual !== undefined)
       ])
     ).run({ seed: 'rc-rule-determinism', effort: 800 , artifacts: { dir: '', persist: 'never' }});
 
diff --git a/tests/rigor/snapshot-parity.test.js b/tests/rigor/snapshot-parity.test.js
index 1905b29..72189c3 100644
--- a/tests/rigor/snapshot-parity.test.js
+++ b/tests/rigor/snapshot-parity.test.js
@@ -123,7 +123,7 @@ describe('Condensed snapshot round trip (rigor)', () => {
         ))
       ],
       rigor.crucible([
-        rigor.invariant('snapshot-parity', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('snapshot-parity', ({ actual }) => actual !== undefined)
       ])
     ).run({ effort: 400, seed: 'snapshot-parity' , artifacts: { dir: '', persist: 'never' }});
 
@@ -175,7 +175,7 @@ describe('Condensed snapshot round trip (rigor)', () => {
         ))
       ],
       rigor.crucible([
-        rigor.invariant('readonly-enforced', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('readonly-enforced', ({ actual }) => actual !== undefined)
       ])
     ).run({ effort: 300, seed: 'snapshot-readonly' , artifacts: { dir: '', persist: 'never' }});
 
diff --git a/tests/rigor/traversal-parity.test.js b/tests/rigor/traversal-parity.test.js
index 0a5403b..4db98e3 100644
--- a/tests/rigor/traversal-parity.test.js
+++ b/tests/rigor/traversal-parity.test.js
@@ -186,7 +186,7 @@ describe('Traversal semantics parity (rigor)', () => {
         ))
       ],
       rigor.crucible([
-        rigor.invariant('chain-direction-parity', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('chain-direction-parity', ({ actual }) => actual !== undefined)
       ])
     ).run({ effort: 1500, seed: 'traversal-chain-direction' , artifacts: { dir: '', persist: 'never' }});
 
@@ -262,7 +262,7 @@ describe('Traversal semantics parity (rigor)', () => {
         ))
       ],
       rigor.crucible([
-        rigor.invariant('ttu-multi-tuple-parity', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('ttu-multi-tuple-parity', ({ actual }) => actual !== undefined)
       ])
     ).run({ effort: 1200, seed: 'traversal-ttu-multituple' , artifacts: { dir: '', persist: 'never' }});
 
@@ -338,7 +338,7 @@ describe('Traversal semantics parity (rigor)', () => {
         ))
       ],
       rigor.crucible([
-        rigor.invariant('update-path-parity', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('update-path-parity', ({ actual }) => actual !== undefined)
       ])
     ).run({ effort: 600, seed: 'traversal-update-path' , artifacts: { dir: '', persist: 'never' }});
 
diff --git a/tests/rigor/tuple-to-userset-rule.test.js b/tests/rigor/tuple-to-userset-rule.test.js
index 4ed9216..e8b6bfc 100644
--- a/tests/rigor/tuple-to-userset-rule.test.js
+++ b/tests/rigor/tuple-to-userset-rule.test.js
@@ -90,7 +90,7 @@ describe('TupleToUsersetRule evaluation (rigor)', () => {
         )
       )],
       rigor.crucible([
-        rigor.invariant('no-tuples', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('no-tuples', ({ actual }) => actual !== undefined)
       ])
     ).run({ effort: 500, seed: 'ttu-no-tuples', artifacts: { dir: '', persist: 'never' }});
 
@@ -139,7 +139,7 @@ describe('TupleToUsersetRule evaluation (rigor)', () => {
         )
       )],
       rigor.crucible([
-        rigor.invariant('min-fusion', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('min-fusion', ({ actual }) => actual !== undefined)
       ])
     ).run({ effort: 800, seed: 'ttu-min-fusion', artifacts: { dir: '', persist: 'never' }});
 
@@ -192,7 +192,7 @@ describe('TupleToUsersetRule evaluation (rigor)', () => {
         )
       )],
       rigor.crucible([
-        rigor.invariant('multi-tuple-max', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('multi-tuple-max', ({ actual }) => actual !== undefined)
       ])
     ).run({ seed: 'ttu-multi-tuple-max', effort: 1500 , artifacts: { dir: '', persist: 'never' }});
 
@@ -230,7 +230,7 @@ describe('TupleToUsersetRule evaluation (rigor)', () => {
         )
       )],
       rigor.crucible([
-        rigor.invariant('possibility-bounded', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('possibility-bounded', ({ actual }) => actual !== undefined)
       ])
     ).run({ seed: 'ttu-possibility-bounded', effort: 800 , artifacts: { dir: '', persist: 'never' }});
 
@@ -290,7 +290,7 @@ describe('TupleToUsersetRule evaluation (rigor)', () => {
     const report = await rigor.campaign(
       [rigor.fn('check', check, rigor.args())],
       rigor.crucible([
-        rigor.invariant('early-exit', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('early-exit', ({ actual }) => actual !== undefined)
       ])
     ).run({ seed: 'ttu-early-exit', effort: 200 , artifacts: { dir: '', persist: 'never' }});
 
@@ -340,7 +340,7 @@ describe('TupleToUsersetRule evaluation (rigor)', () => {
     const report = await rigor.campaign(
       [rigor.fn('check', check, rigor.args())],
       rigor.crucible([
-        rigor.invariant('cycle-detection', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('cycle-detection', ({ actual }) => actual !== undefined)
       ])
     ).run({ seed: 'ttu-cycle-detection', effort: 200 , artifacts: { dir: '', persist: 'never' }});
 
diff --git a/tests/rigor/validity-parity.test.js b/tests/rigor/validity-parity.test.js
index 394bd9b..8ca21a7 100644
--- a/tests/rigor/validity-parity.test.js
+++ b/tests/rigor/validity-parity.test.js
@@ -138,7 +138,7 @@ describe('Possibilistic validity metadata (rigor)', () => {
         };
       }, rigor.args(rigor.gen.array(rigor.gen.oneOf(['finite_sample', 'anytime', 'conformal', 'approximate', 'heuristic', 'unknown']), 1, 4)))],
       rigor.crucible([
-        rigor.invariant('helper invariants', ({ error, errorMessage, actual }) => !error && !errorMessage && Object.values(actual).every(Boolean))
+        rigor.invariant('helper invariants', ({ actual }) => !!actual && Object.values(actual).every(Boolean))
       ])
     ).run({ effort: 200, seed: 'validity-helpers-2026', artifacts: { dir: '', persist: 'never' } });
 
diff --git a/tests/rigor/zanzibar-consistency.test.js b/tests/rigor/zanzibar-consistency.test.js
index e25023b..0f2d652 100644
--- a/tests/rigor/zanzibar-consistency.test.js
+++ b/tests/rigor/zanzibar-consistency.test.js
@@ -91,7 +91,7 @@ describe('Authorization state consistency (rigor)', () => {
         ))
       ],
       rigor.crucible([
-        rigor.invariant('ttu-mutation', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('ttu-mutation', ({ actual }) => actual !== undefined)
       ])
     ).run({ effort: 400, seed: 'consistency-ttu-mutation' , artifacts: { dir: '', persist: 'never' }});
 
@@ -152,7 +152,7 @@ describe('Authorization state consistency (rigor)', () => {
         ))
       ],
       rigor.crucible([
-        rigor.invariant('chain-mutation', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('chain-mutation', ({ actual }) => actual !== undefined)
       ])
     ).run({ effort: 400, seed: 'consistency-chain-mutation' , artifacts: { dir: '', persist: 'never' }});
 
@@ -220,7 +220,7 @@ describe('Authorization state consistency (rigor)', () => {
         ))
       ],
       rigor.crucible([
-        rigor.invariant('config-change', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('config-change', ({ actual }) => actual !== undefined)
       ])
     ).run({ effort: 400, seed: 'consistency-config-change' , artifacts: { dir: '', persist: 'never' }});
 
@@ -272,7 +272,7 @@ describe('Authorization state consistency (rigor)', () => {
         ))
       ],
       rigor.crucible([
-        rigor.invariant('threshold-excludes', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('threshold-excludes', ({ actual }) => actual !== undefined)
       ])
     ).run({ effort: 400, seed: 'consistency-threshold' , artifacts: { dir: '', persist: 'never' }});
 
@@ -338,7 +338,7 @@ describe('Authorization state consistency (rigor)', () => {
         ))
       ],
       rigor.crucible([
-        rigor.invariant('values-complete', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('values-complete', ({ actual }) => actual !== undefined)
       ])
     ).run({ effort: 400, seed: 'consistency-value-completeness' , artifacts: { dir: '', persist: 'never' }});
 
@@ -375,7 +375,7 @@ describe('Authorization state consistency (rigor)', () => {
         ))
       ],
       rigor.crucible([
-        rigor.invariant('deterministic', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('deterministic', ({ actual }) => actual !== undefined)
       ])
     ).run({ effort: 300, seed: 'consistency-determinism' , artifacts: { dir: '', persist: 'never' }});
 
diff --git a/tests/rigor/zanzibar-defeasible-dsl-comparator.test.js b/tests/rigor/zanzibar-defeasible-dsl-comparator.test.js
index e1d227b..ae13b6e 100644
--- a/tests/rigor/zanzibar-defeasible-dsl-comparator.test.js
+++ b/tests/rigor/zanzibar-defeasible-dsl-comparator.test.js
@@ -98,7 +98,7 @@ describe('Defeasible logic, DSL parity, aggregation (rigor)', () => {
         ))
       ],
       rigor.crucible([
-        rigor.invariant('defeasible-semantics', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('defeasible-semantics', ({ actual }) => actual !== undefined)
       ])
     ).run({ effort: 600, seed: 'defeasible-semantics' , artifacts: { dir: '', persist: 'never' }});
 
@@ -172,7 +172,7 @@ describe('Defeasible logic, DSL parity, aggregation (rigor)', () => {
         ))
       ],
       rigor.crucible([
-        rigor.invariant('dsl-parity', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('dsl-parity', ({ actual }) => actual !== undefined)
       ])
     ).run({ effort: 400, seed: 'dsl-runtime-parity' , artifacts: { dir: '', persist: 'never' }});
 
@@ -263,7 +263,7 @@ describe('Defeasible logic, DSL parity, aggregation (rigor)', () => {
         ))
       ],
       rigor.crucible([
-        rigor.invariant('aggregation-complete', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('aggregation-complete', ({ actual }) => actual !== undefined)
       ])
     ).run({ effort: 500, seed: 'aggregation-completeness' , artifacts: { dir: '', persist: 'never' }});
 
diff --git a/tests/rigor/zanzibar-semantics.test.js b/tests/rigor/zanzibar-semantics.test.js
index f7e2d34..cdc2091 100644
--- a/tests/rigor/zanzibar-semantics.test.js
+++ b/tests/rigor/zanzibar-semantics.test.js
@@ -310,7 +310,7 @@ describe('Zanzibar rewrite-rule semantics (rigor)', () => {
         ))
       ],
       rigor.crucible([
-        rigor.invariant('union-max', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('union-max', ({ actual }) => actual !== undefined)
       ])
     ).run({ effort: 400, seed: 'zanzibar-union' , artifacts: { dir: '', persist: 'never' }});
 
@@ -354,7 +354,7 @@ describe('Zanzibar rewrite-rule semantics (rigor)', () => {
         ))
       ],
       rigor.crucible([
-        rigor.invariant('intersection-min', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('intersection-min', ({ actual }) => actual !== undefined)
       ])
     ).run({ effort: 400, seed: 'zanzibar-intersection' , artifacts: { dir: '', persist: 'never' }});
 
@@ -397,7 +397,7 @@ describe('Zanzibar rewrite-rule semantics (rigor)', () => {
         ))
       ],
       rigor.crucible([
-        rigor.invariant('exclusion-blocks', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('exclusion-blocks', ({ actual }) => actual !== undefined)
       ])
     ).run({ effort: 400, seed: 'zanzibar-exclusion' , artifacts: { dir: '', persist: 'never' }});
 
@@ -432,7 +432,7 @@ describe('Zanzibar rewrite-rule semantics (rigor)', () => {
         ))
       ],
       rigor.crucible([
-        rigor.invariant('expand-parity', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('expand-parity', ({ actual }) => actual !== undefined)
       ])
     ).run({ effort: 800, seed: 'zanzibar-expand-parity' , artifacts: { dir: '', persist: 'never' }});
 
@@ -488,7 +488,7 @@ describe('Zanzibar rewrite-rule semantics (rigor)', () => {
         ))
       ],
       rigor.crucible([
-        rigor.invariant('nested-weakest-link', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('nested-weakest-link', ({ actual }) => actual !== undefined)
       ])
     ).run({ effort: 500, seed: 'zanzibar-nested-chain' , artifacts: { dir: '', persist: 'never' }});
 
@@ -544,7 +544,7 @@ describe('Zanzibar rewrite-rule semantics (rigor)', () => {
         ))
       ],
       rigor.crucible([
-        rigor.invariant('cycle-safe', ({ error, errorMessage }) => !error && !errorMessage)
+        rigor.invariant('cycle-safe', ({ actual }) => actual !== undefined)
       ])
     ).run({ effort: 400, seed: 'zanzibar-cycle-safety' , artifacts: { dir: '', persist: 'never' }});
 
diff --git a/tests/rules/chain-condition-step.test.js b/tests/rules/chain-condition-step.test.js
index ceaf977..9d7a14c 100644
--- a/tests/rules/chain-condition-step.test.js
+++ b/tests/rules/chain-condition-step.test.js
@@ -81,14 +81,29 @@ describe('ChainRule condition step (rule-based final hop)', () => {
     assert.equal(res.reason, 'no_chain_path_found');
   });
 
-  it('rejects a condition step that is not the final step', () => {
+  it('expands an intermediate condition step (rule-based reachability)', () => {
+    arbiter.setRelationConfig('member_of', { type: 'direct' });
+    arbiter.addNode('group:g', 'group');
+    arbiter.addRelation('user:u', 'member_of', 'group:g', { possibility: 1.0 });
+    arbiter.addRelation('group:g', 'can_view', 'doc:d', { possibility: 0.8 });
+    // [condition(member_of unless banned), can_view] — the condition step is
+    // INTERMEDIATE and discovers its reachable nodes (its base relation's
+    // neighbors from the source, filtered by its defeater).
+    const intermediateConfig = {
+      type: 'logical',
+      when: { intersection: { rules: [{ type: 'direct', relation: 'member_of' }], aggregator: 'min' } },
+      unless: { union: { rules: [{ type: 'direct', relation: 'banned', _subjectIsObject: true }], aggregator: 'max' } }
+    };
     const rule = {
       type: 'chain',
-      steps: [{ rule: CONDITION_CONFIG, conditionStep: true }, 'can_view']
+      steps: [{ rule: intermediateConfig, conditionStep: true }, 'can_view']
     };
-    const res = evalRule('user:u', 'doc:d', rule);
-    assert.equal(res.possibility, 0);
-    assert.equal(res.reason, 'condition_step_not_final');
+    // g not banned → reachable via condition → can_view → doc
+    assert.ok(Math.abs(evalRule('user:u', 'doc:d', rule).possibility - 0.8) < 1e-9);
+    // banning g filters it out of the intermediate expansion → no path
+    arbiter.addRelation('group:g', 'banned', 'group:g', { possibility: 1.0 });
+    const denied = evalRule('user:u', 'doc:d', rule);
+    assert.equal(denied.possibility, 0);
   });
 
   it('combines across multiple parallel intermediates (max aggregation)', () => {