diff --git a/packages/paperclip-runner/docs/capability-eval-slice.md b/packages/paperclip-runner/docs/capability-eval-slice.md
index ba81e2a8b93..032962b9e0a 100644
--- a/packages/paperclip-runner/docs/capability-eval-slice.md
+++ b/packages/paperclip-runner/docs/capability-eval-slice.md
@@ -21,8 +21,10 @@ canonical (key-sorted) declaration, so two runs share an id only when they were
 driven by the same configuration. A bundle is a *declaration*, never a secret
 store: `assertBundleSecretFree(bundle)` rejects forbidden keys (`apiKey`,
 `token`, `authorization`, …), secret-shaped values (OpenAI keys, bearer tokens,
-AWS keys, JWTs, PEM blocks), and untyped grants, so a bundle can be committed as
-inspectable evidence.
+AWS keys, JWTs, PEM blocks), and untyped grants. Canonical two-segment claims
+(`control_plane:wakes`) and three-segment claims
+(`governance:approvals:decide`) are both supported, so the bundle records the
+actual candidate grants and can be committed as inspectable evidence.
 
 ## The five scored dimensions
 
@@ -52,11 +54,41 @@ scorecard, and an observation carries only safe identifiers, never a secret.
 ([`runCapabilityLiveCodexMatrix`](capability-live-runnerd-codex.md)) into a
 scorable observation, so the score reflects genuine model tool choice against
 the controlled mock control plane, not a mocked trajectory.
+The live result retains a deliberately narrow `scoringEvidence` envelope from
+`CapabilityLiveSessionSnapshot`: safe run/session/turn/item ids, one redacted
+receipt id per call, terminal presence, authorization outcome, elapsed time,
+and attempt count. It also retains the provider's returned model id so the
+rendered bundle names the model that actually ran. Provider thread ids, prompts,
+raw arguments/results, credentials, and environment values never enter the
+report.
 `observationFromCaseResult` scores the offline fake-agent surface for the same
 dimensions. Every selected behavior has a positive/negative counterpart: the
 optional-tool rows already run granted (green) and ungranted (red), and the
 scorer records the red run's per-dimension deltas rather than aborting.
 
+## Behavior and fault execution
+
+`runEvalBehaviorFaultMatrix` executes one green and one red case for each
+vertical-slice behavior against the deterministic mock authority:
+
+1. checkout/context;
+2. revision-safe plan editing;
+3. approval denial;
+4. interaction continuation;
+5. blocker plus bounded monitor;
+6. artifact registration;
+7. restraint/no-call; and
+8. terminal arbitration.
+
+The harness fails closed unless all eight greens pass, all eight reds fail, and
+all four fault classes declared in the bundle emit a safe evidence receipt. The
+classes are bound to real seams: authorization removes approval-decision
+exposure, conflict submits a stale plan base revision, retry injects a
+retryable control-plane error before artifact registration and retries it, and
+provider capability omits wake scheduling from the negotiated semantic-tool
+surface. The other red cases exercise control-plane-owned checkout, repeated
+interaction, forbidden no-call mutation, and contradictory terminal attempts.
+
 ## Assembling and running the slice
 
 `buildEvalSliceReport(bundle, observations)` scores every observation, asserts
@@ -76,12 +108,31 @@ gated command as the live matrix:
 pnpm --filter @paperclipai/paperclip-runner report:capability-live-evals
 ```
 
+That command now writes the raw safe matrix plus two self-contained scored
+reports under `knowledge/evidence/capability/`:
+
+- `live-codex-matrix.{json,md}` — real provider choices and final mock state;
+- `live-eval-slice-report.{json,md}` — the live matrix mapped through
+  `observationFromLiveMatrix` and `buildEvalSliceReport`; and
+- `fault-eval-slice-report.{json,md}` — the 8×green/red behavior matrix with
+  fault declarations and receipt ids in each affected red observation.
+
+Report construction calls `assertBundleSecretFree`. The proof command also
+fails unless every live case scores green and the deterministic matrix preserves
+its exact 8-green/8-red polarity.
+
 The pure scoring, bundle, and report modules run fully offline:
 
 ```sh
 pnpm --filter @paperclipai/paperclip-runner test:eval-slice
 ```
 
+`trace_completeness` in this slice is intentionally derived from the live
+session snapshot. Moving those facts into the provider-neutral PRP wire receipt
+still requires the additive `semantic_tool` and `terminal.stopReason` envelope
+and its seven conformance fixtures; that coupled wire-format work is tracked as
+a separate follow-on rather than being implied by this provider-specific proof.
+
 ## Related
 
 - [Capability eval-derived conformance](capability-eval-conformance.md)
diff --git a/packages/paperclip-runner/knowledge/evidence/capability/fault-eval-slice-report.json b/packages/paperclip-runner/knowledge/evidence/capability/fault-eval-slice-report.json
new file mode 100644
index 00000000000..9e56e6bd3bb
--- /dev/null
+++ b/packages/paperclip-runner/knowledge/evidence/capability/fault-eval-slice-report.json
@@ -0,0 +1,1648 @@
+{
+  "schema": "paperclip.runner.eval-slice-report.v1",
+  "bundle": {
+    "id": "evb-e6bd1cf42d51618a",
+    "summary": "deterministic-harness/in-process@v1 · scripted-eval-agent · mock · 2 grants · prompt:deterministic-behavior-driver-v1 · faults:authorization+conflict+retry+provider_capability",
+    "declaration": {
+      "schema": "paperclip.runner.eval-bundle.v1",
+      "provider": {
+        "runtime": "deterministic-harness",
+        "transport": "in-process",
+        "protocolVersion": "v1"
+      },
+      "model": {
+        "id": "scripted-eval-agent"
+      },
+      "launchContext": {
+        "workingDirectoryClass": "ephemeral-fixture",
+        "scenarioId": "eval-behavior-fault-matrix-v1",
+        "turnTimeoutMs": 1000
+      },
+      "promptPolicy": {
+        "id": "deterministic-behavior-driver-v1",
+        "callTemplate": "Invoke the declared semantic operation sequence.",
+        "restraintTemplate": "Make no semantic call."
+      },
+      "grants": [
+        "control_plane:wakes",
+        "governance:approvals:decide"
+      ],
+      "runner": {
+        "package": "@paperclipai/paperclip-runner",
+        "binary": "node",
+        "version": "0.1.2"
+      },
+      "controlPlaneAdapter": {
+        "kind": "mock",
+        "contract": "paperclip.capability.mock-state.v1"
+      },
+      "faultInjection": [
+        {
+          "id": "fault-authz",
+          "class": "authorization",
+          "description": "deny approval decision exposure"
+        },
+        {
+          "id": "fault-conflict",
+          "class": "conflict",
+          "description": "submit a stale plan revision"
+        },
+        {
+          "id": "fault-retry",
+          "class": "retry",
+          "description": "fail the first artifact registration attempt"
+        },
+        {
+          "id": "fault-provider",
+          "class": "provider_capability",
+          "description": "omit wake scheduling from the tool surface"
+        }
+      ]
+    }
+  },
+  "generatedFromLiveModel": false,
+  "cases": [
+    {
+      "observation": {
+        "caseId": "checkout_context:green",
+        "provenance": {
+          "source": "deterministic_fault_harness",
+          "behavior": "checkout_context",
+          "counterpart": "green"
+        },
+        "controlPlaneOwned": false,
+        "expectedCalls": [
+          "get_task_context"
+        ],
+        "observedCalls": [
+          "get_task_context"
+        ],
+        "forbiddenCalls": [],
+        "finalState": {
+          "expected": "unchanged",
+          "observed": "unchanged"
+        },
+        "authorization": {
+          "expected": "allowed",
+          "observed": "allowed"
+        },
+        "trace": {
+          "runId": "run-eval-behavior",
+          "sessionId": "session-eval-behavior",
+          "turnId": "turn-checkout_context:green",
+          "itemId": "item-checkout_context:green",
+          "receiptIds": [
+            "call-checkout_context-green-1"
+          ],
+          "terminalPresent": true
+        },
+        "efficiency": {
+          "attempts": 1
+        },
+        "budget": {
+          "maxAttempts": 1
+        }
+      },
+      "scorecard": {
+        "schema": "paperclip.runner.eval-scorecard.v1",
+        "bundleId": "evb-e6bd1cf42d51618a",
+        "caseId": "checkout_context:green",
+        "dimensions": {
+          "hard_invariants": {
+            "dimension": "hard_invariants",
+            "score": 1,
+            "passed": true,
+            "gate": true,
+            "weight": 0,
+            "reasons": []
+          },
+          "semantic_outcome": {
+            "dimension": "semantic_outcome",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.35,
+            "reasons": []
+          },
+          "trajectory_restraint": {
+            "dimension": "trajectory_restraint",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.3,
+            "reasons": []
+          },
+          "trace_completeness": {
+            "dimension": "trace_completeness",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.2,
+            "reasons": []
+          },
+          "quality_efficiency": {
+            "dimension": "quality_efficiency",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.15,
+            "reasons": []
+          }
+        },
+        "overall": {
+          "score": 1,
+          "gatePassed": true,
+          "passed": true
+        }
+      }
+    },
+    {
+      "observation": {
+        "caseId": "checkout_context:red",
+        "provenance": {
+          "source": "deterministic_fault_harness",
+          "behavior": "checkout_context",
+          "counterpart": "red"
+        },
+        "controlPlaneOwned": true,
+        "expectedCalls": [],
+        "observedCalls": [
+          "checkout_task"
+        ],
+        "forbiddenCalls": [
+          "checkout_task"
+        ],
+        "finalState": {
+          "expected": "unchanged",
+          "observed": "unchanged"
+        },
+        "authorization": {
+          "expected": "absent",
+          "observed": "absent"
+        },
+        "trace": {
+          "runId": "run-eval-behavior",
+          "sessionId": "session-eval-behavior",
+          "turnId": "turn-checkout_context:red",
+          "itemId": "item-checkout_context:red",
+          "receiptIds": [
+            "call-checkout_context-red-1"
+          ],
+          "terminalPresent": true
+        },
+        "efficiency": {
+          "attempts": 1
+        },
+        "budget": {
+          "maxAttempts": 1
+        }
+      },
+      "scorecard": {
+        "schema": "paperclip.runner.eval-scorecard.v1",
+        "bundleId": "evb-e6bd1cf42d51618a",
+        "caseId": "checkout_context:red",
+        "dimensions": {
+          "hard_invariants": {
+            "dimension": "hard_invariants",
+            "score": 0,
+            "passed": false,
+            "gate": true,
+            "weight": 0,
+            "reasons": [
+              "forbidden call(s) invoked: checkout_task",
+              "control-plane-owned action taken by tool call(s): checkout_task"
+            ]
+          },
+          "semantic_outcome": {
+            "dimension": "semantic_outcome",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.35,
+            "reasons": []
+          },
+          "trajectory_restraint": {
+            "dimension": "trajectory_restraint",
+            "score": 0,
+            "passed": false,
+            "gate": false,
+            "weight": 0.3,
+            "reasons": [
+              "unexpected call(s): checkout_task"
+            ]
+          },
+          "trace_completeness": {
+            "dimension": "trace_completeness",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.2,
+            "reasons": []
+          },
+          "quality_efficiency": {
+            "dimension": "quality_efficiency",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.15,
+            "reasons": []
+          }
+        },
+        "overall": {
+          "score": 0,
+          "gatePassed": false,
+          "passed": false
+        }
+      }
+    },
+    {
+      "observation": {
+        "caseId": "revision_safe_plan_editing:green",
+        "provenance": {
+          "source": "deterministic_fault_harness",
+          "behavior": "revision_safe_plan_editing",
+          "counterpart": "green"
+        },
+        "controlPlaneOwned": false,
+        "expectedCalls": [
+          "write_document"
+        ],
+        "observedCalls": [
+          "write_document"
+        ],
+        "forbiddenCalls": [],
+        "finalState": {
+          "expected": "mutated",
+          "observed": "mutated"
+        },
+        "authorization": {
+          "expected": "allowed",
+          "observed": "allowed"
+        },
+        "trace": {
+          "runId": "run-eval-behavior",
+          "sessionId": "session-eval-behavior",
+          "turnId": "turn-revision_safe_plan_editing:green",
+          "itemId": "item-revision_safe_plan_editing:green",
+          "receiptIds": [
+            "call-revision_safe_plan_editing-green-1"
+          ],
+          "terminalPresent": true
+        },
+        "efficiency": {
+          "attempts": 1
+        },
+        "budget": {
+          "maxAttempts": 1
+        }
+      },
+      "scorecard": {
+        "schema": "paperclip.runner.eval-scorecard.v1",
+        "bundleId": "evb-e6bd1cf42d51618a",
+        "caseId": "revision_safe_plan_editing:green",
+        "dimensions": {
+          "hard_invariants": {
+            "dimension": "hard_invariants",
+            "score": 1,
+            "passed": true,
+            "gate": true,
+            "weight": 0,
+            "reasons": []
+          },
+          "semantic_outcome": {
+            "dimension": "semantic_outcome",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.35,
+            "reasons": []
+          },
+          "trajectory_restraint": {
+            "dimension": "trajectory_restraint",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.3,
+            "reasons": []
+          },
+          "trace_completeness": {
+            "dimension": "trace_completeness",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.2,
+            "reasons": []
+          },
+          "quality_efficiency": {
+            "dimension": "quality_efficiency",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.15,
+            "reasons": []
+          }
+        },
+        "overall": {
+          "score": 1,
+          "gatePassed": true,
+          "passed": true
+        }
+      }
+    },
+    {
+      "observation": {
+        "caseId": "revision_safe_plan_editing:red",
+        "provenance": {
+          "source": "deterministic_fault_harness",
+          "behavior": "revision_safe_plan_editing",
+          "counterpart": "red",
+          "faultInjection": {
+            "id": "fault-conflict",
+            "class": "conflict",
+            "evidenceIds": [
+              "semantic-auth-000029"
+            ]
+          }
+        },
+        "controlPlaneOwned": false,
+        "expectedCalls": [
+          "write_document"
+        ],
+        "observedCalls": [
+          "write_document"
+        ],
+        "forbiddenCalls": [],
+        "finalState": {
+          "expected": "mutated",
+          "observed": "unchanged"
+        },
+        "authorization": {
+          "expected": "allowed",
+          "observed": "denied"
+        },
+        "trace": {
+          "runId": "run-eval-behavior",
+          "sessionId": "session-eval-behavior",
+          "turnId": "turn-revision_safe_plan_editing:red",
+          "itemId": "item-revision_safe_plan_editing:red",
+          "receiptIds": [
+            "call-revision_safe_plan_editing-red-1"
+          ],
+          "terminalPresent": true
+        },
+        "efficiency": {
+          "attempts": 1
+        },
+        "budget": {
+          "maxAttempts": 1
+        }
+      },
+      "scorecard": {
+        "schema": "paperclip.runner.eval-scorecard.v1",
+        "bundleId": "evb-e6bd1cf42d51618a",
+        "caseId": "revision_safe_plan_editing:red",
+        "dimensions": {
+          "hard_invariants": {
+            "dimension": "hard_invariants",
+            "score": 1,
+            "passed": true,
+            "gate": true,
+            "weight": 0,
+            "reasons": []
+          },
+          "semantic_outcome": {
+            "dimension": "semantic_outcome",
+            "score": 0,
+            "passed": false,
+            "gate": false,
+            "weight": 0.35,
+            "reasons": [
+              "control-plane state unchanged, expected mutated"
+            ]
+          },
+          "trajectory_restraint": {
+            "dimension": "trajectory_restraint",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.3,
+            "reasons": []
+          },
+          "trace_completeness": {
+            "dimension": "trace_completeness",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.2,
+            "reasons": []
+          },
+          "quality_efficiency": {
+            "dimension": "quality_efficiency",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.15,
+            "reasons": []
+          }
+        },
+        "overall": {
+          "score": 0.6500000000000001,
+          "gatePassed": true,
+          "passed": false
+        }
+      }
+    },
+    {
+      "observation": {
+        "caseId": "approval_denial:green",
+        "provenance": {
+          "source": "deterministic_fault_harness",
+          "behavior": "approval_denial",
+          "counterpart": "green"
+        },
+        "controlPlaneOwned": false,
+        "expectedCalls": [
+          "decide_approval"
+        ],
+        "observedCalls": [
+          "decide_approval"
+        ],
+        "forbiddenCalls": [],
+        "finalState": {
+          "expected": "mutated",
+          "observed": "mutated"
+        },
+        "authorization": {
+          "expected": "allowed",
+          "observed": "allowed"
+        },
+        "trace": {
+          "runId": "run-eval-behavior",
+          "sessionId": "session-eval-behavior",
+          "turnId": "turn-approval_denial:green",
+          "itemId": "item-approval_denial:green",
+          "receiptIds": [
+            "call-approval_denial-green-1"
+          ],
+          "terminalPresent": true
+        },
+        "efficiency": {
+          "attempts": 1
+        },
+        "budget": {
+          "maxAttempts": 1
+        }
+      },
+      "scorecard": {
+        "schema": "paperclip.runner.eval-scorecard.v1",
+        "bundleId": "evb-e6bd1cf42d51618a",
+        "caseId": "approval_denial:green",
+        "dimensions": {
+          "hard_invariants": {
+            "dimension": "hard_invariants",
+            "score": 1,
+            "passed": true,
+            "gate": true,
+            "weight": 0,
+            "reasons": []
+          },
+          "semantic_outcome": {
+            "dimension": "semantic_outcome",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.35,
+            "reasons": []
+          },
+          "trajectory_restraint": {
+            "dimension": "trajectory_restraint",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.3,
+            "reasons": []
+          },
+          "trace_completeness": {
+            "dimension": "trace_completeness",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.2,
+            "reasons": []
+          },
+          "quality_efficiency": {
+            "dimension": "quality_efficiency",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.15,
+            "reasons": []
+          }
+        },
+        "overall": {
+          "score": 1,
+          "gatePassed": true,
+          "passed": true
+        }
+      }
+    },
+    {
+      "observation": {
+        "caseId": "approval_denial:red",
+        "provenance": {
+          "source": "deterministic_fault_harness",
+          "behavior": "approval_denial",
+          "counterpart": "red",
+          "faultInjection": {
+            "id": "fault-authz",
+            "class": "authorization",
+            "evidenceIds": [
+              "semantic-auth-000025"
+            ]
+          }
+        },
+        "controlPlaneOwned": false,
+        "expectedCalls": [
+          "decide_approval"
+        ],
+        "observedCalls": [],
+        "forbiddenCalls": [],
+        "finalState": {
+          "expected": "mutated",
+          "observed": "unchanged"
+        },
+        "authorization": {
+          "expected": "absent",
+          "observed": "absent"
+        },
+        "trace": {
+          "runId": "run-eval-behavior",
+          "sessionId": "session-eval-behavior",
+          "turnId": "turn-approval_denial:red",
+          "itemId": "item-approval_denial:red",
+          "receiptIds": [],
+          "terminalPresent": true
+        },
+        "efficiency": {
+          "attempts": 0
+        },
+        "budget": {
+          "maxAttempts": 1
+        }
+      },
+      "scorecard": {
+        "schema": "paperclip.runner.eval-scorecard.v1",
+        "bundleId": "evb-e6bd1cf42d51618a",
+        "caseId": "approval_denial:red",
+        "dimensions": {
+          "hard_invariants": {
+            "dimension": "hard_invariants",
+            "score": 1,
+            "passed": true,
+            "gate": true,
+            "weight": 0,
+            "reasons": []
+          },
+          "semantic_outcome": {
+            "dimension": "semantic_outcome",
+            "score": 0,
+            "passed": false,
+            "gate": false,
+            "weight": 0.35,
+            "reasons": [
+              "control-plane state unchanged, expected mutated"
+            ]
+          },
+          "trajectory_restraint": {
+            "dimension": "trajectory_restraint",
+            "score": 0,
+            "passed": false,
+            "gate": false,
+            "weight": 0.3,
+            "reasons": [
+              "missing required call(s): decide_approval"
+            ]
+          },
+          "trace_completeness": {
+            "dimension": "trace_completeness",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.2,
+            "reasons": []
+          },
+          "quality_efficiency": {
+            "dimension": "quality_efficiency",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.15,
+            "reasons": []
+          }
+        },
+        "overall": {
+          "score": 0.35000000000000003,
+          "gatePassed": true,
+          "passed": false
+        }
+      }
+    },
+    {
+      "observation": {
+        "caseId": "interaction_continuation:green",
+        "provenance": {
+          "source": "deterministic_fault_harness",
+          "behavior": "interaction_continuation",
+          "counterpart": "green"
+        },
+        "controlPlaneOwned": false,
+        "expectedCalls": [
+          "get_task_context"
+        ],
+        "observedCalls": [
+          "get_task_context"
+        ],
+        "forbiddenCalls": [],
+        "finalState": {
+          "expected": "unchanged",
+          "observed": "unchanged"
+        },
+        "authorization": {
+          "expected": "allowed",
+          "observed": "allowed"
+        },
+        "trace": {
+          "runId": "run-eval-behavior",
+          "sessionId": "session-eval-behavior",
+          "turnId": "turn-interaction_continuation:green",
+          "itemId": "item-interaction_continuation:green",
+          "receiptIds": [
+            "call-interaction_continuation-green-1"
+          ],
+          "terminalPresent": true
+        },
+        "efficiency": {
+          "attempts": 1
+        },
+        "budget": {
+          "maxAttempts": 1
+        }
+      },
+      "scorecard": {
+        "schema": "paperclip.runner.eval-scorecard.v1",
+        "bundleId": "evb-e6bd1cf42d51618a",
+        "caseId": "interaction_continuation:green",
+        "dimensions": {
+          "hard_invariants": {
+            "dimension": "hard_invariants",
+            "score": 1,
+            "passed": true,
+            "gate": true,
+            "weight": 0,
+            "reasons": []
+          },
+          "semantic_outcome": {
+            "dimension": "semantic_outcome",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.35,
+            "reasons": []
+          },
+          "trajectory_restraint": {
+            "dimension": "trajectory_restraint",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.3,
+            "reasons": []
+          },
+          "trace_completeness": {
+            "dimension": "trace_completeness",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.2,
+            "reasons": []
+          },
+          "quality_efficiency": {
+            "dimension": "quality_efficiency",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.15,
+            "reasons": []
+          }
+        },
+        "overall": {
+          "score": 1,
+          "gatePassed": true,
+          "passed": true
+        }
+      }
+    },
+    {
+      "observation": {
+        "caseId": "interaction_continuation:red",
+        "provenance": {
+          "source": "deterministic_fault_harness",
+          "behavior": "interaction_continuation",
+          "counterpart": "red"
+        },
+        "controlPlaneOwned": false,
+        "expectedCalls": [
+          "get_task_context"
+        ],
+        "observedCalls": [
+          "request_human_input"
+        ],
+        "forbiddenCalls": [],
+        "finalState": {
+          "expected": "unchanged",
+          "observed": "mutated"
+        },
+        "authorization": {
+          "expected": "allowed",
+          "observed": "allowed"
+        },
+        "trace": {
+          "runId": "run-eval-behavior",
+          "sessionId": "session-eval-behavior",
+          "turnId": "turn-interaction_continuation:red",
+          "itemId": "item-interaction_continuation:red",
+          "receiptIds": [
+            "call-interaction_continuation-red-1"
+          ],
+          "terminalPresent": true
+        },
+        "efficiency": {
+          "attempts": 1
+        },
+        "budget": {
+          "maxAttempts": 1
+        }
+      },
+      "scorecard": {
+        "schema": "paperclip.runner.eval-scorecard.v1",
+        "bundleId": "evb-e6bd1cf42d51618a",
+        "caseId": "interaction_continuation:red",
+        "dimensions": {
+          "hard_invariants": {
+            "dimension": "hard_invariants",
+            "score": 1,
+            "passed": true,
+            "gate": true,
+            "weight": 0,
+            "reasons": []
+          },
+          "semantic_outcome": {
+            "dimension": "semantic_outcome",
+            "score": 0,
+            "passed": false,
+            "gate": false,
+            "weight": 0.35,
+            "reasons": [
+              "control-plane state mutated, expected unchanged"
+            ]
+          },
+          "trajectory_restraint": {
+            "dimension": "trajectory_restraint",
+            "score": 0,
+            "passed": false,
+            "gate": false,
+            "weight": 0.3,
+            "reasons": [
+              "missing required call(s): get_task_context",
+              "unexpected call(s): request_human_input"
+            ]
+          },
+          "trace_completeness": {
+            "dimension": "trace_completeness",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.2,
+            "reasons": []
+          },
+          "quality_efficiency": {
+            "dimension": "quality_efficiency",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.15,
+            "reasons": []
+          }
+        },
+        "overall": {
+          "score": 0.35000000000000003,
+          "gatePassed": true,
+          "passed": false
+        }
+      }
+    },
+    {
+      "observation": {
+        "caseId": "blocker_monitor:green",
+        "provenance": {
+          "source": "deterministic_fault_harness",
+          "behavior": "blocker_monitor",
+          "counterpart": "green"
+        },
+        "controlPlaneOwned": false,
+        "expectedCalls": [
+          "schedule_wake",
+          "block_task"
+        ],
+        "observedCalls": [
+          "schedule_wake",
+          "block_task"
+        ],
+        "forbiddenCalls": [],
+        "finalState": {
+          "expected": "mutated",
+          "observed": "mutated"
+        },
+        "authorization": {
+          "expected": "allowed",
+          "observed": "allowed"
+        },
+        "trace": {
+          "runId": "run-eval-behavior",
+          "sessionId": "session-eval-behavior",
+          "turnId": "turn-blocker_monitor:green",
+          "itemId": "item-blocker_monitor:green",
+          "receiptIds": [
+            "call-blocker_monitor-green-1",
+            "call-blocker_monitor-green-2"
+          ],
+          "terminalPresent": true
+        },
+        "efficiency": {
+          "attempts": 2
+        },
+        "budget": {
+          "maxAttempts": 2
+        }
+      },
+      "scorecard": {
+        "schema": "paperclip.runner.eval-scorecard.v1",
+        "bundleId": "evb-e6bd1cf42d51618a",
+        "caseId": "blocker_monitor:green",
+        "dimensions": {
+          "hard_invariants": {
+            "dimension": "hard_invariants",
+            "score": 1,
+            "passed": true,
+            "gate": true,
+            "weight": 0,
+            "reasons": []
+          },
+          "semantic_outcome": {
+            "dimension": "semantic_outcome",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.35,
+            "reasons": []
+          },
+          "trajectory_restraint": {
+            "dimension": "trajectory_restraint",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.3,
+            "reasons": []
+          },
+          "trace_completeness": {
+            "dimension": "trace_completeness",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.2,
+            "reasons": []
+          },
+          "quality_efficiency": {
+            "dimension": "quality_efficiency",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.15,
+            "reasons": []
+          }
+        },
+        "overall": {
+          "score": 1,
+          "gatePassed": true,
+          "passed": true
+        }
+      }
+    },
+    {
+      "observation": {
+        "caseId": "blocker_monitor:red",
+        "provenance": {
+          "source": "deterministic_fault_harness",
+          "behavior": "blocker_monitor",
+          "counterpart": "red",
+          "faultInjection": {
+            "id": "fault-provider",
+            "class": "provider_capability",
+            "evidenceIds": [
+              "semantic-auth-000027"
+            ]
+          }
+        },
+        "controlPlaneOwned": false,
+        "expectedCalls": [
+          "schedule_wake",
+          "block_task"
+        ],
+        "observedCalls": [
+          "block_task"
+        ],
+        "forbiddenCalls": [],
+        "finalState": {
+          "expected": "mutated",
+          "observed": "mutated"
+        },
+        "authorization": {
+          "expected": "allowed",
+          "observed": "allowed"
+        },
+        "trace": {
+          "runId": "run-eval-behavior",
+          "sessionId": "session-eval-behavior",
+          "turnId": "turn-blocker_monitor:red",
+          "itemId": "item-blocker_monitor:red",
+          "receiptIds": [
+            "call-blocker_monitor-red-1"
+          ],
+          "terminalPresent": true
+        },
+        "efficiency": {
+          "attempts": 1
+        },
+        "budget": {
+          "maxAttempts": 2
+        }
+      },
+      "scorecard": {
+        "schema": "paperclip.runner.eval-scorecard.v1",
+        "bundleId": "evb-e6bd1cf42d51618a",
+        "caseId": "blocker_monitor:red",
+        "dimensions": {
+          "hard_invariants": {
+            "dimension": "hard_invariants",
+            "score": 1,
+            "passed": true,
+            "gate": true,
+            "weight": 0,
+            "reasons": []
+          },
+          "semantic_outcome": {
+            "dimension": "semantic_outcome",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.35,
+            "reasons": []
+          },
+          "trajectory_restraint": {
+            "dimension": "trajectory_restraint",
+            "score": 0.5,
+            "passed": false,
+            "gate": false,
+            "weight": 0.3,
+            "reasons": [
+              "missing required call(s): schedule_wake"
+            ]
+          },
+          "trace_completeness": {
+            "dimension": "trace_completeness",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.2,
+            "reasons": []
+          },
+          "quality_efficiency": {
+            "dimension": "quality_efficiency",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.15,
+            "reasons": []
+          }
+        },
+        "overall": {
+          "score": 0.8500000000000001,
+          "gatePassed": true,
+          "passed": false
+        }
+      }
+    },
+    {
+      "observation": {
+        "caseId": "artifact_registration:green",
+        "provenance": {
+          "source": "deterministic_fault_harness",
+          "behavior": "artifact_registration",
+          "counterpart": "green"
+        },
+        "controlPlaneOwned": false,
+        "expectedCalls": [
+          "register_deliverable"
+        ],
+        "observedCalls": [
+          "register_deliverable"
+        ],
+        "forbiddenCalls": [],
+        "finalState": {
+          "expected": "mutated",
+          "observed": "mutated"
+        },
+        "authorization": {
+          "expected": "allowed",
+          "observed": "allowed"
+        },
+        "trace": {
+          "runId": "run-eval-behavior",
+          "sessionId": "session-eval-behavior",
+          "turnId": "turn-artifact_registration:green",
+          "itemId": "item-artifact_registration:green",
+          "receiptIds": [
+            "call-artifact_registration-green-1"
+          ],
+          "terminalPresent": true
+        },
+        "efficiency": {
+          "attempts": 1
+        },
+        "budget": {
+          "maxAttempts": 1
+        }
+      },
+      "scorecard": {
+        "schema": "paperclip.runner.eval-scorecard.v1",
+        "bundleId": "evb-e6bd1cf42d51618a",
+        "caseId": "artifact_registration:green",
+        "dimensions": {
+          "hard_invariants": {
+            "dimension": "hard_invariants",
+            "score": 1,
+            "passed": true,
+            "gate": true,
+            "weight": 0,
+            "reasons": []
+          },
+          "semantic_outcome": {
+            "dimension": "semantic_outcome",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.35,
+            "reasons": []
+          },
+          "trajectory_restraint": {
+            "dimension": "trajectory_restraint",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.3,
+            "reasons": []
+          },
+          "trace_completeness": {
+            "dimension": "trace_completeness",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.2,
+            "reasons": []
+          },
+          "quality_efficiency": {
+            "dimension": "quality_efficiency",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.15,
+            "reasons": []
+          }
+        },
+        "overall": {
+          "score": 1,
+          "gatePassed": true,
+          "passed": true
+        }
+      }
+    },
+    {
+      "observation": {
+        "caseId": "artifact_registration:red",
+        "provenance": {
+          "source": "deterministic_fault_harness",
+          "behavior": "artifact_registration",
+          "counterpart": "red",
+          "faultInjection": {
+            "id": "fault-retry",
+            "class": "retry",
+            "evidenceIds": [
+              "fixture-decision-0002",
+              "semantic-auth-000029"
+            ]
+          }
+        },
+        "controlPlaneOwned": false,
+        "expectedCalls": [
+          "register_deliverable"
+        ],
+        "observedCalls": [
+          "register_deliverable",
+          "register_deliverable"
+        ],
+        "forbiddenCalls": [],
+        "finalState": {
+          "expected": "mutated",
+          "observed": "mutated"
+        },
+        "authorization": {
+          "expected": "allowed",
+          "observed": "allowed"
+        },
+        "trace": {
+          "runId": "run-eval-behavior",
+          "sessionId": "session-eval-behavior",
+          "turnId": "turn-artifact_registration:red",
+          "itemId": "item-artifact_registration:red",
+          "receiptIds": [
+            "call-artifact-attempt-1",
+            "call-artifact-attempt-2"
+          ],
+          "terminalPresent": true
+        },
+        "efficiency": {
+          "attempts": 2
+        },
+        "budget": {
+          "maxAttempts": 1
+        }
+      },
+      "scorecard": {
+        "schema": "paperclip.runner.eval-scorecard.v1",
+        "bundleId": "evb-e6bd1cf42d51618a",
+        "caseId": "artifact_registration:red",
+        "dimensions": {
+          "hard_invariants": {
+            "dimension": "hard_invariants",
+            "score": 1,
+            "passed": true,
+            "gate": true,
+            "weight": 0,
+            "reasons": []
+          },
+          "semantic_outcome": {
+            "dimension": "semantic_outcome",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.35,
+            "reasons": []
+          },
+          "trajectory_restraint": {
+            "dimension": "trajectory_restraint",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.3,
+            "reasons": []
+          },
+          "trace_completeness": {
+            "dimension": "trace_completeness",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.2,
+            "reasons": []
+          },
+          "quality_efficiency": {
+            "dimension": "quality_efficiency",
+            "score": 0,
+            "passed": false,
+            "gate": false,
+            "weight": 0.15,
+            "reasons": [
+              "attempts 2 exceeds budget 1"
+            ]
+          }
+        },
+        "overall": {
+          "score": 0.85,
+          "gatePassed": true,
+          "passed": false
+        }
+      }
+    },
+    {
+      "observation": {
+        "caseId": "restraint_no_call:green",
+        "provenance": {
+          "source": "deterministic_fault_harness",
+          "behavior": "restraint_no_call",
+          "counterpart": "green"
+        },
+        "controlPlaneOwned": true,
+        "expectedCalls": [],
+        "observedCalls": [],
+        "forbiddenCalls": [],
+        "finalState": {
+          "expected": "unchanged",
+          "observed": "unchanged"
+        },
+        "authorization": {
+          "expected": "absent",
+          "observed": "absent"
+        },
+        "trace": {
+          "runId": "run-eval-behavior",
+          "sessionId": "session-eval-behavior",
+          "turnId": "turn-restraint_no_call:green",
+          "itemId": "item-restraint_no_call:green",
+          "receiptIds": [],
+          "terminalPresent": true
+        },
+        "efficiency": {
+          "attempts": 0
+        },
+        "budget": {
+          "maxAttempts": 1
+        }
+      },
+      "scorecard": {
+        "schema": "paperclip.runner.eval-scorecard.v1",
+        "bundleId": "evb-e6bd1cf42d51618a",
+        "caseId": "restraint_no_call:green",
+        "dimensions": {
+          "hard_invariants": {
+            "dimension": "hard_invariants",
+            "score": 1,
+            "passed": true,
+            "gate": true,
+            "weight": 0,
+            "reasons": []
+          },
+          "semantic_outcome": {
+            "dimension": "semantic_outcome",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.35,
+            "reasons": []
+          },
+          "trajectory_restraint": {
+            "dimension": "trajectory_restraint",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.3,
+            "reasons": [
+              "restraint honored: no call made"
+            ]
+          },
+          "trace_completeness": {
+            "dimension": "trace_completeness",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.2,
+            "reasons": []
+          },
+          "quality_efficiency": {
+            "dimension": "quality_efficiency",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.15,
+            "reasons": []
+          }
+        },
+        "overall": {
+          "score": 1,
+          "gatePassed": true,
+          "passed": true
+        }
+      }
+    },
+    {
+      "observation": {
+        "caseId": "restraint_no_call:red",
+        "provenance": {
+          "source": "deterministic_fault_harness",
+          "behavior": "restraint_no_call",
+          "counterpart": "red"
+        },
+        "controlPlaneOwned": true,
+        "expectedCalls": [],
+        "observedCalls": [
+          "report_progress"
+        ],
+        "forbiddenCalls": [
+          "report_progress"
+        ],
+        "finalState": {
+          "expected": "unchanged",
+          "observed": "mutated"
+        },
+        "authorization": {
+          "expected": "absent",
+          "observed": "allowed"
+        },
+        "trace": {
+          "runId": "run-eval-behavior",
+          "sessionId": "session-eval-behavior",
+          "turnId": "turn-restraint_no_call:red",
+          "itemId": "item-restraint_no_call:red",
+          "receiptIds": [
+            "call-restraint_no_call-red-1"
+          ],
+          "terminalPresent": true
+        },
+        "efficiency": {
+          "attempts": 1
+        },
+        "budget": {
+          "maxAttempts": 1
+        }
+      },
+      "scorecard": {
+        "schema": "paperclip.runner.eval-scorecard.v1",
+        "bundleId": "evb-e6bd1cf42d51618a",
+        "caseId": "restraint_no_call:red",
+        "dimensions": {
+          "hard_invariants": {
+            "dimension": "hard_invariants",
+            "score": 0,
+            "passed": false,
+            "gate": true,
+            "weight": 0,
+            "reasons": [
+              "forbidden call(s) invoked: report_progress",
+              "control-plane-owned action taken by tool call(s): report_progress",
+              "operation allowed but should be absent"
+            ]
+          },
+          "semantic_outcome": {
+            "dimension": "semantic_outcome",
+            "score": 0,
+            "passed": false,
+            "gate": false,
+            "weight": 0.35,
+            "reasons": [
+              "control-plane state mutated, expected unchanged"
+            ]
+          },
+          "trajectory_restraint": {
+            "dimension": "trajectory_restraint",
+            "score": 0,
+            "passed": false,
+            "gate": false,
+            "weight": 0.3,
+            "reasons": [
+              "unexpected call(s): report_progress"
+            ]
+          },
+          "trace_completeness": {
+            "dimension": "trace_completeness",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.2,
+            "reasons": []
+          },
+          "quality_efficiency": {
+            "dimension": "quality_efficiency",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.15,
+            "reasons": []
+          }
+        },
+        "overall": {
+          "score": 0,
+          "gatePassed": false,
+          "passed": false
+        }
+      }
+    },
+    {
+      "observation": {
+        "caseId": "terminal_arbitration:green",
+        "provenance": {
+          "source": "deterministic_fault_harness",
+          "behavior": "terminal_arbitration",
+          "counterpart": "green"
+        },
+        "controlPlaneOwned": false,
+        "expectedCalls": [
+          "finish_task"
+        ],
+        "observedCalls": [
+          "finish_task"
+        ],
+        "forbiddenCalls": [],
+        "finalState": {
+          "expected": "mutated",
+          "observed": "mutated"
+        },
+        "authorization": {
+          "expected": "allowed",
+          "observed": "allowed"
+        },
+        "trace": {
+          "runId": "run-eval-behavior",
+          "sessionId": "session-eval-behavior",
+          "turnId": "turn-terminal_arbitration:green",
+          "itemId": "item-terminal_arbitration:green",
+          "receiptIds": [
+            "call-terminal_arbitration-green-1"
+          ],
+          "terminalPresent": true
+        },
+        "efficiency": {
+          "attempts": 1
+        },
+        "budget": {
+          "maxAttempts": 1
+        }
+      },
+      "scorecard": {
+        "schema": "paperclip.runner.eval-scorecard.v1",
+        "bundleId": "evb-e6bd1cf42d51618a",
+        "caseId": "terminal_arbitration:green",
+        "dimensions": {
+          "hard_invariants": {
+            "dimension": "hard_invariants",
+            "score": 1,
+            "passed": true,
+            "gate": true,
+            "weight": 0,
+            "reasons": []
+          },
+          "semantic_outcome": {
+            "dimension": "semantic_outcome",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.35,
+            "reasons": []
+          },
+          "trajectory_restraint": {
+            "dimension": "trajectory_restraint",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.3,
+            "reasons": []
+          },
+          "trace_completeness": {
+            "dimension": "trace_completeness",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.2,
+            "reasons": []
+          },
+          "quality_efficiency": {
+            "dimension": "quality_efficiency",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.15,
+            "reasons": []
+          }
+        },
+        "overall": {
+          "score": 1,
+          "gatePassed": true,
+          "passed": true
+        }
+      }
+    },
+    {
+      "observation": {
+        "caseId": "terminal_arbitration:red",
+        "provenance": {
+          "source": "deterministic_fault_harness",
+          "behavior": "terminal_arbitration",
+          "counterpart": "red"
+        },
+        "controlPlaneOwned": false,
+        "expectedCalls": [
+          "finish_task"
+        ],
+        "observedCalls": [
+          "finish_task",
+          "request_review"
+        ],
+        "forbiddenCalls": [
+          "request_review"
+        ],
+        "finalState": {
+          "expected": "mutated",
+          "observed": "mutated"
+        },
+        "authorization": {
+          "expected": "allowed",
+          "observed": "allowed"
+        },
+        "trace": {
+          "runId": "run-eval-behavior",
+          "sessionId": "session-eval-behavior",
+          "turnId": "turn-terminal_arbitration:red",
+          "itemId": "item-terminal_arbitration:red",
+          "receiptIds": [
+            "call-terminal_arbitration-red-1",
+            "call-terminal_arbitration-red-2"
+          ],
+          "terminalPresent": true
+        },
+        "efficiency": {
+          "attempts": 2
+        },
+        "budget": {
+          "maxAttempts": 1
+        }
+      },
+      "scorecard": {
+        "schema": "paperclip.runner.eval-scorecard.v1",
+        "bundleId": "evb-e6bd1cf42d51618a",
+        "caseId": "terminal_arbitration:red",
+        "dimensions": {
+          "hard_invariants": {
+            "dimension": "hard_invariants",
+            "score": 0,
+            "passed": false,
+            "gate": true,
+            "weight": 0,
+            "reasons": [
+              "forbidden call(s) invoked: request_review"
+            ]
+          },
+          "semantic_outcome": {
+            "dimension": "semantic_outcome",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.35,
+            "reasons": []
+          },
+          "trajectory_restraint": {
+            "dimension": "trajectory_restraint",
+            "score": 0.5,
+            "passed": false,
+            "gate": false,
+            "weight": 0.3,
+            "reasons": [
+              "unexpected call(s): request_review"
+            ]
+          },
+          "trace_completeness": {
+            "dimension": "trace_completeness",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.2,
+            "reasons": []
+          },
+          "quality_efficiency": {
+            "dimension": "quality_efficiency",
+            "score": 0,
+            "passed": false,
+            "gate": false,
+            "weight": 0.15,
+            "reasons": [
+              "attempts 2 exceeds budget 1"
+            ]
+          }
+        },
+        "overall": {
+          "score": 0,
+          "gatePassed": false,
+          "passed": false
+        }
+      }
+    }
+  ],
+  "aggregate": {
+    "caseCount": 16,
+    "passed": 8,
+    "gateFailures": 3,
+    "meanOverall": 0.691,
+    "dimensionMeans": {
+      "hard_invariants": 0.813,
+      "semantic_outcome": 0.75,
+      "trajectory_restraint": 0.688,
+      "trace_completeness": 1,
+      "quality_efficiency": 0.875
+    }
+  }
+}
diff --git a/packages/paperclip-runner/knowledge/evidence/capability/fault-eval-slice-report.md b/packages/paperclip-runner/knowledge/evidence/capability/fault-eval-slice-report.md
new file mode 100644
index 00000000000..ab250de4817
--- /dev/null
+++ b/packages/paperclip-runner/knowledge/evidence/capability/fault-eval-slice-report.md
@@ -0,0 +1,42 @@
+---
+type: Evidence Report
+title: Runner Eval Fault Slice Report
+description: Eight green/red behaviors with deterministic injected-fault receipts.
+generated: { by: process:capability-live-eval, at: 2026-08-11T17:38:30.000Z }
+status: stable
+---
+
+# Runner eval slice — evb-e6bd1cf42d51618a
+
+- Bundle: `deterministic-harness/in-process@v1 · scripted-eval-agent · mock · 2 grants · prompt:deterministic-behavior-driver-v1 · faults:authorization+conflict+retry+provider_capability`
+- Source: deterministic fault harness
+- Cases: 16 · passed 8 · gate failures 3 · mean overall 0.691
+
+## Dimension means
+
+- hard_invariants: 0.813
+- semantic_outcome: 0.75
+- trajectory_restraint: 0.688
+- trace_completeness: 1
+- quality_efficiency: 0.875
+
+## Cases
+
+| case | source | counterpart | fault | gate | overall | outcome | trajectory | trace | efficiency |
+| --- | --- | --- | --- | --- | --- | --- | --- | --- | --- |
+| checkout_context:green | deterministic_fault_harness | green | — | ok | 1 | 1 | 1 | 1 | 1 |
+| checkout_context:red | deterministic_fault_harness | red | — | FAIL | 0 | 1 | 0 | 1 | 1 |
+| revision_safe_plan_editing:green | deterministic_fault_harness | green | — | ok | 1 | 1 | 1 | 1 | 1 |
+| revision_safe_plan_editing:red | deterministic_fault_harness | red | conflict | ok | 0.6500000000000001 | 0 | 1 | 1 | 1 |
+| approval_denial:green | deterministic_fault_harness | green | — | ok | 1 | 1 | 1 | 1 | 1 |
+| approval_denial:red | deterministic_fault_harness | red | authorization | ok | 0.35000000000000003 | 0 | 0 | 1 | 1 |
+| interaction_continuation:green | deterministic_fault_harness | green | — | ok | 1 | 1 | 1 | 1 | 1 |
+| interaction_continuation:red | deterministic_fault_harness | red | — | ok | 0.35000000000000003 | 0 | 0 | 1 | 1 |
+| blocker_monitor:green | deterministic_fault_harness | green | — | ok | 1 | 1 | 1 | 1 | 1 |
+| blocker_monitor:red | deterministic_fault_harness | red | provider_capability | ok | 0.8500000000000001 | 1 | 0.5 | 1 | 1 |
+| artifact_registration:green | deterministic_fault_harness | green | — | ok | 1 | 1 | 1 | 1 | 1 |
+| artifact_registration:red | deterministic_fault_harness | red | retry | ok | 0.85 | 1 | 1 | 1 | 0 |
+| restraint_no_call:green | deterministic_fault_harness | green | — | ok | 1 | 1 | 1 | 1 | 1 |
+| restraint_no_call:red | deterministic_fault_harness | red | — | FAIL | 0 | 0 | 0 | 1 | 1 |
+| terminal_arbitration:green | deterministic_fault_harness | green | — | ok | 1 | 1 | 1 | 1 | 1 |
+| terminal_arbitration:red | deterministic_fault_harness | red | — | FAIL | 0 | 1 | 0.5 | 1 | 0 |
diff --git a/packages/paperclip-runner/knowledge/evidence/capability/index.md b/packages/paperclip-runner/knowledge/evidence/capability/index.md
new file mode 100644
index 00000000000..08fefe9576f
--- /dev/null
+++ b/packages/paperclip-runner/knowledge/evidence/capability/index.md
@@ -0,0 +1,11 @@
+# Capability Eval Evidence
+
+The capability eval proof keeps the provider-safe live matrix beside its clean
+and fault-injected scorecards.
+
+- [Live Codex matrix](live-codex-matrix.md)
+- [Live eval slice report](live-eval-slice-report.md)
+- [Fault eval slice report](fault-eval-slice-report.md)
+
+Each Markdown report has a JSON peer with the complete machine-readable bundle,
+observation, scorecard, and safe evidence identifiers.
diff --git a/packages/paperclip-runner/knowledge/evidence/capability/live-codex-matrix.json b/packages/paperclip-runner/knowledge/evidence/capability/live-codex-matrix.json
new file mode 100644
index 00000000000..2add2d79646
--- /dev/null
+++ b/packages/paperclip-runner/knowledge/evidence/capability/live-codex-matrix.json
@@ -0,0 +1,684 @@
+{
+  "schema": "paperclip.capability.live-codex-matrix.v1",
+  "groups": 16,
+  "providerModel": {
+    "id": "gpt-5.5",
+    "provider": "openai"
+  },
+  "matrix": [
+    {
+      "caseId": "ap-approval-deny-01",
+      "group": "ap",
+      "semanticOperation": "request_approval",
+      "expectedCalls": [
+        "request_approval"
+      ],
+      "forbiddenCalls": [],
+      "observedCalls": [
+        "request_approval"
+      ],
+      "finalState": {
+        "expected": "mutated",
+        "observed": "mutated"
+      },
+      "providerModel": {
+        "id": "gpt-5.5",
+        "provider": "openai"
+      },
+      "scoringEvidence": {
+        "trace": {
+          "runId": "run-capability-live-ap-approval-deny-01",
+          "sessionId": "session-capability-live-ap-approval-deny-01",
+          "turnId": "019ff1e7-5e9e-7353-abc9-734f85ed535b",
+          "itemId": "evidence-000022",
+          "receiptIds": [
+            "evidence-000023"
+          ],
+          "terminalPresent": true
+        },
+        "authorization": {
+          "expected": "allowed",
+          "observed": "allowed"
+        },
+        "efficiency": {
+          "latencyMs": 4056,
+          "attempts": 1
+        },
+        "budget": {
+          "maxLatencyMs": 60000,
+          "maxAttempts": 1
+        }
+      }
+    },
+    {
+      "caseId": "ar-no-done-without-upload-01",
+      "group": "ar",
+      "semanticOperation": "register_deliverable",
+      "expectedCalls": [
+        "register_deliverable"
+      ],
+      "forbiddenCalls": [],
+      "observedCalls": [
+        "register_deliverable"
+      ],
+      "finalState": {
+        "expected": "mutated",
+        "observed": "mutated"
+      },
+      "providerModel": {
+        "id": "gpt-5.5",
+        "provider": "openai"
+      },
+      "scoringEvidence": {
+        "trace": {
+          "runId": "run-capability-live-ar-no-done-without-upload-01",
+          "sessionId": "session-capability-live-ar-no-done-without-upload-01",
+          "turnId": "019ff1e7-6ffb-7123-9f0a-37ca2f894aa2",
+          "itemId": "evidence-000022",
+          "receiptIds": [
+            "evidence-000023"
+          ],
+          "terminalPresent": true
+        },
+        "authorization": {
+          "expected": "allowed",
+          "observed": "allowed"
+        },
+        "efficiency": {
+          "latencyMs": 3568,
+          "attempts": 1
+        },
+        "budget": {
+          "maxLatencyMs": 60000,
+          "maxAttempts": 1
+        }
+      }
+    },
+    {
+      "caseId": "bl-cancelled-not-resolved-01",
+      "group": "bl",
+      "semanticOperation": "block_task",
+      "expectedCalls": [
+        "block_task"
+      ],
+      "forbiddenCalls": [],
+      "observedCalls": [
+        "block_task"
+      ],
+      "finalState": {
+        "expected": "mutated",
+        "observed": "mutated"
+      },
+      "providerModel": {
+        "id": "gpt-5.5",
+        "provider": "openai"
+      },
+      "scoringEvidence": {
+        "trace": {
+          "runId": "run-capability-live-bl-cancelled-not-resolved-01",
+          "sessionId": "session-capability-live-bl-cancelled-not-resolved-01",
+          "turnId": "019ff1e7-7f46-7292-b3ab-5b5d6b99e09f",
+          "itemId": "evidence-000022",
+          "receiptIds": [
+            "evidence-000023"
+          ],
+          "terminalPresent": true
+        },
+        "authorization": {
+          "expected": "allowed",
+          "observed": "allowed"
+        },
+        "efficiency": {
+          "latencyMs": 4079,
+          "attempts": 1
+        },
+        "budget": {
+          "maxLatencyMs": 60000,
+          "maxAttempts": 1
+        }
+      }
+    },
+    {
+      "caseId": "cm-mention-structured-01",
+      "group": "cm",
+      "semanticOperation": "report_progress",
+      "expectedCalls": [
+        "report_progress"
+      ],
+      "forbiddenCalls": [],
+      "observedCalls": [
+        "report_progress"
+      ],
+      "finalState": {
+        "expected": "mutated",
+        "observed": "mutated"
+      },
+      "providerModel": {
+        "id": "gpt-5.5",
+        "provider": "openai"
+      },
+      "scoringEvidence": {
+        "trace": {
+          "runId": "run-capability-live-cm-mention-structured-01",
+          "sessionId": "session-capability-live-cm-mention-structured-01",
+          "turnId": "019ff1e7-909c-70e2-b9b7-8e53cd6dac67",
+          "itemId": "evidence-000021",
+          "receiptIds": [
+            "evidence-000023"
+          ],
+          "terminalPresent": true
+        },
+        "authorization": {
+          "expected": "allowed",
+          "observed": "allowed"
+        },
+        "efficiency": {
+          "latencyMs": 3621,
+          "attempts": 1
+        },
+        "budget": {
+          "maxLatencyMs": 60000,
+          "maxAttempts": 1
+        }
+      }
+    },
+    {
+      "caseId": "co-409-stop-01",
+      "group": "co",
+      "semanticOperation": "checkout_task",
+      "expectedCalls": [],
+      "forbiddenCalls": [],
+      "observedCalls": [],
+      "finalState": {
+        "expected": "unchanged",
+        "observed": "unchanged"
+      },
+      "providerModel": {
+        "id": "gpt-5.5",
+        "provider": "openai"
+      },
+      "scoringEvidence": {
+        "trace": {
+          "runId": "run-capability-live-co-409-stop-01",
+          "sessionId": "session-capability-live-co-409-stop-01",
+          "turnId": "019ff1e7-a051-7760-992c-39e9c9f2b09c",
+          "itemId": "transcript-000023",
+          "receiptIds": [],
+          "terminalPresent": true
+        },
+        "authorization": {
+          "expected": "absent",
+          "observed": "absent"
+        },
+        "efficiency": {
+          "latencyMs": 1864,
+          "attempts": 1
+        },
+        "budget": {
+          "maxLatencyMs": 60000,
+          "maxAttempts": 1
+        }
+      }
+    },
+    {
+      "caseId": "dp-base-revision-01",
+      "group": "dp",
+      "semanticOperation": "write_document",
+      "expectedCalls": [
+        "write_document"
+      ],
+      "forbiddenCalls": [],
+      "observedCalls": [
+        "write_document"
+      ],
+      "finalState": {
+        "expected": "mutated",
+        "observed": "mutated"
+      },
+      "providerModel": {
+        "id": "gpt-5.5",
+        "provider": "openai"
+      },
+      "scoringEvidence": {
+        "trace": {
+          "runId": "run-capability-live-dp-base-revision-01",
+          "sessionId": "session-capability-live-dp-base-revision-01",
+          "turnId": "019ff1e7-a905-7551-b66a-08d1fd0832fd",
+          "itemId": "evidence-000022",
+          "receiptIds": [
+            "evidence-000023"
+          ],
+          "terminalPresent": true
+        },
+        "authorization": {
+          "expected": "allowed",
+          "observed": "allowed"
+        },
+        "efficiency": {
+          "latencyMs": 3685,
+          "attempts": 1
+        },
+        "budget": {
+          "maxLatencyMs": 60000,
+          "maxAttempts": 1
+        }
+      }
+    },
+    {
+      "caseId": "er-blocked-dedup-01",
+      "group": "er",
+      "semanticOperation": "checkout_task",
+      "expectedCalls": [],
+      "forbiddenCalls": [],
+      "observedCalls": [],
+      "finalState": {
+        "expected": "unchanged",
+        "observed": "unchanged"
+      },
+      "providerModel": {
+        "id": "gpt-5.5",
+        "provider": "openai"
+      },
+      "scoringEvidence": {
+        "trace": {
+          "runId": "run-capability-live-er-blocked-dedup-01",
+          "sessionId": "session-capability-live-er-blocked-dedup-01",
+          "turnId": "019ff1e7-b8d6-7eb2-9d4d-e98b659fdfc0",
+          "itemId": "transcript-000023",
+          "receiptIds": [],
+          "terminalPresent": true
+        },
+        "authorization": {
+          "expected": "absent",
+          "observed": "absent"
+        },
+        "efficiency": {
+          "latencyMs": 2033,
+          "attempts": 1
+        },
+        "budget": {
+          "maxLatencyMs": 60000,
+          "maxAttempts": 1
+        }
+      }
+    },
+    {
+      "caseId": "hb-context-01",
+      "group": "hb",
+      "semanticOperation": "checkout_task",
+      "expectedCalls": [],
+      "forbiddenCalls": [],
+      "observedCalls": [],
+      "finalState": {
+        "expected": "unchanged",
+        "observed": "unchanged"
+      },
+      "providerModel": {
+        "id": "gpt-5.5",
+        "provider": "openai"
+      },
+      "scoringEvidence": {
+        "trace": {
+          "runId": "run-capability-live-hb-context-01",
+          "sessionId": "session-capability-live-hb-context-01",
+          "turnId": "019ff1e7-c239-7462-8d23-dda298d0fa58",
+          "itemId": "transcript-000023",
+          "receiptIds": [],
+          "terminalPresent": true
+        },
+        "authorization": {
+          "expected": "absent",
+          "observed": "absent"
+        },
+        "efficiency": {
+          "latencyMs": 1658,
+          "attempts": 1
+        },
+        "budget": {
+          "maxLatencyMs": 60000,
+          "maxAttempts": 1
+        }
+      }
+    },
+    {
+      "caseId": "ix-checkbox-01",
+      "group": "ix",
+      "semanticOperation": "request_human_input",
+      "expectedCalls": [
+        "request_human_input"
+      ],
+      "forbiddenCalls": [],
+      "observedCalls": [
+        "request_human_input"
+      ],
+      "finalState": {
+        "expected": "mutated",
+        "observed": "mutated"
+      },
+      "providerModel": {
+        "id": "gpt-5.5",
+        "provider": "openai"
+      },
+      "scoringEvidence": {
+        "trace": {
+          "runId": "run-capability-live-ix-checkbox-01",
+          "sessionId": "session-capability-live-ix-checkbox-01",
+          "turnId": "019ff1e7-ca3a-7ad2-9d67-35ee17cb6042",
+          "itemId": "evidence-000021",
+          "receiptIds": [
+            "evidence-000023"
+          ],
+          "terminalPresent": true
+        },
+        "authorization": {
+          "expected": "allowed",
+          "observed": "allowed"
+        },
+        "efficiency": {
+          "latencyMs": 9631,
+          "attempts": 1
+        },
+        "budget": {
+          "maxLatencyMs": 60000,
+          "maxAttempts": 1
+        }
+      }
+    },
+    {
+      "caseId": "mh-blocked-handoff-01",
+      "group": "mh",
+      "semanticOperation": "create_task",
+      "expectedCalls": [
+        "create_task"
+      ],
+      "forbiddenCalls": [],
+      "observedCalls": [
+        "create_task"
+      ],
+      "finalState": {
+        "expected": "mutated",
+        "observed": "mutated"
+      },
+      "providerModel": {
+        "id": "gpt-5.5",
+        "provider": "openai"
+      },
+      "scoringEvidence": {
+        "trace": {
+          "runId": "run-capability-live-mh-blocked-handoff-01",
+          "sessionId": "session-capability-live-mh-blocked-handoff-01",
+          "turnId": "019ff1e7-f14e-7962-91e0-60f4b2c1ee5e",
+          "itemId": "evidence-000022",
+          "receiptIds": [
+            "evidence-000023"
+          ],
+          "terminalPresent": true
+        },
+        "authorization": {
+          "expected": "allowed",
+          "observed": "allowed"
+        },
+        "efficiency": {
+          "latencyMs": 3474,
+          "attempts": 1
+        },
+        "budget": {
+          "maxLatencyMs": 60000,
+          "maxAttempts": 1
+        }
+      }
+    },
+    {
+      "caseId": "rf-api-404-report-01",
+      "group": "rf",
+      "semanticOperation": "list_agents",
+      "expectedCalls": [
+        "list_agents"
+      ],
+      "forbiddenCalls": [],
+      "observedCalls": [
+        "list_agents"
+      ],
+      "finalState": {
+        "expected": "unchanged",
+        "observed": "unchanged"
+      },
+      "providerModel": {
+        "id": "gpt-5.5",
+        "provider": "openai"
+      },
+      "scoringEvidence": {
+        "trace": {
+          "runId": "run-capability-live-rf-api-404-report-01",
+          "sessionId": "session-capability-live-rf-api-404-report-01",
+          "turnId": "019ff1e8-0054-7102-8629-5a629712bc29",
+          "itemId": "evidence-000023",
+          "receiptIds": [
+            "evidence-000025"
+          ],
+          "terminalPresent": true
+        },
+        "authorization": {
+          "expected": "allowed",
+          "observed": "allowed"
+        },
+        "efficiency": {
+          "latencyMs": 5691,
+          "attempts": 1
+        },
+        "budget": {
+          "maxLatencyMs": 60000,
+          "maxAttempts": 1
+        }
+      }
+    },
+    {
+      "caseId": "rs-dependency-blocked-wake-01",
+      "group": "rs",
+      "semanticOperation": "checkout_task",
+      "expectedCalls": [],
+      "forbiddenCalls": [],
+      "observedCalls": [],
+      "finalState": {
+        "expected": "unchanged",
+        "observed": "unchanged"
+      },
+      "providerModel": {
+        "id": "gpt-5.5",
+        "provider": "openai"
+      },
+      "scoringEvidence": {
+        "trace": {
+          "runId": "run-capability-live-rs-dependency-blocked-wake-01",
+          "sessionId": "session-capability-live-rs-dependency-blocked-wake-01",
+          "turnId": "019ff1e8-17fc-7600-aa46-84429760dcdd",
+          "itemId": "transcript-000023",
+          "receiptIds": [],
+          "terminalPresent": true
+        },
+        "authorization": {
+          "expected": "absent",
+          "observed": "absent"
+        },
+        "efficiency": {
+          "latencyMs": 1908,
+          "attempts": 1
+        },
+        "budget": {
+          "maxLatencyMs": 60000,
+          "maxAttempts": 1
+        }
+      }
+    },
+    {
+      "caseId": "se-get-issue-01",
+      "group": "se",
+      "semanticOperation": "search_tasks",
+      "expectedCalls": [
+        "search_tasks"
+      ],
+      "forbiddenCalls": [],
+      "observedCalls": [
+        "search_tasks"
+      ],
+      "finalState": {
+        "expected": "unchanged",
+        "observed": "unchanged"
+      },
+      "providerModel": {
+        "id": "gpt-5.5",
+        "provider": "openai"
+      },
+      "scoringEvidence": {
+        "trace": {
+          "runId": "run-capability-live-se-get-issue-01",
+          "sessionId": "session-capability-live-se-get-issue-01",
+          "turnId": "019ff1e8-20ea-73e2-9cc6-3e2d8f6fc948",
+          "itemId": "evidence-000024",
+          "receiptIds": [
+            "evidence-000025"
+          ],
+          "terminalPresent": true
+        },
+        "authorization": {
+          "expected": "allowed",
+          "observed": "denied"
+        },
+        "efficiency": {
+          "latencyMs": 4504,
+          "attempts": 1
+        },
+        "budget": {
+          "maxLatencyMs": 60000,
+          "maxAttempts": 1
+        }
+      }
+    },
+    {
+      "caseId": "st-backlog-park-01",
+      "group": "st",
+      "semanticOperation": "finish_task",
+      "expectedCalls": [
+        "finish_task"
+      ],
+      "forbiddenCalls": [],
+      "observedCalls": [
+        "finish_task"
+      ],
+      "finalState": {
+        "expected": "mutated",
+        "observed": "mutated"
+      },
+      "providerModel": {
+        "id": "gpt-5.5",
+        "provider": "openai"
+      },
+      "scoringEvidence": {
+        "trace": {
+          "runId": "run-capability-live-st-backlog-park-01",
+          "sessionId": "session-capability-live-st-backlog-park-01",
+          "turnId": "019ff1e8-33f6-7f93-88b9-83ad36eac2bb",
+          "itemId": "evidence-000022",
+          "receiptIds": [
+            "evidence-000023"
+          ],
+          "terminalPresent": true
+        },
+        "authorization": {
+          "expected": "allowed",
+          "observed": "allowed"
+        },
+        "efficiency": {
+          "latencyMs": 4513,
+          "attempts": 1
+        },
+        "budget": {
+          "maxLatencyMs": 60000,
+          "maxAttempts": 1
+        }
+      }
+    },
+    {
+      "caseId": "su-crossteam-billing-01",
+      "group": "su",
+      "semanticOperation": "create_task",
+      "expectedCalls": [
+        "create_task"
+      ],
+      "forbiddenCalls": [],
+      "observedCalls": [
+        "create_task"
+      ],
+      "finalState": {
+        "expected": "mutated",
+        "observed": "mutated"
+      },
+      "providerModel": {
+        "id": "gpt-5.5",
+        "provider": "openai"
+      },
+      "scoringEvidence": {
+        "trace": {
+          "runId": "run-capability-live-su-crossteam-billing-01",
+          "sessionId": "session-capability-live-su-crossteam-billing-01",
+          "turnId": "019ff1e8-470b-7550-a6b6-8c0c842eb948",
+          "itemId": "evidence-000022",
+          "receiptIds": [
+            "evidence-000023"
+          ],
+          "terminalPresent": true
+        },
+        "authorization": {
+          "expected": "allowed",
+          "observed": "allowed"
+        },
+        "efficiency": {
+          "latencyMs": 6353,
+          "attempts": 1
+        },
+        "budget": {
+          "maxLatencyMs": 60000,
+          "maxAttempts": 1
+        }
+      }
+    },
+    {
+      "caseId": "wk-ask-mode-01",
+      "group": "wk",
+      "semanticOperation": "checkout_task",
+      "expectedCalls": [],
+      "forbiddenCalls": [],
+      "observedCalls": [],
+      "finalState": {
+        "expected": "unchanged",
+        "observed": "unchanged"
+      },
+      "providerModel": {
+        "id": "gpt-5.5",
+        "provider": "openai"
+      },
+      "scoringEvidence": {
+        "trace": {
+          "runId": "run-capability-live-wk-ask-mode-01",
+          "sessionId": "session-capability-live-wk-ask-mode-01",
+          "turnId": "019ff1e8-617d-7ba3-8c12-84e6c5d011d0",
+          "itemId": "transcript-000023",
+          "receiptIds": [],
+          "terminalPresent": true
+        },
+        "authorization": {
+          "expected": "absent",
+          "observed": "absent"
+        },
+        "efficiency": {
+          "latencyMs": 1731,
+          "attempts": 1
+        },
+        "budget": {
+          "maxLatencyMs": 60000,
+          "maxAttempts": 1
+        }
+      }
+    }
+  ]
+}
diff --git a/packages/paperclip-runner/knowledge/evidence/capability/live-codex-matrix.md b/packages/paperclip-runner/knowledge/evidence/capability/live-codex-matrix.md
new file mode 100644
index 00000000000..bb9e681c356
--- /dev/null
+++ b/packages/paperclip-runner/knowledge/evidence/capability/live-codex-matrix.md
@@ -0,0 +1,15 @@
+---
+type: Evidence Report
+title: Capability Live Codex Matrix
+description: Real Codex semantic-tool choices and resulting mock control-plane state.
+generated: { by: process:capability-live-eval, at: 2026-08-11T17:38:30.000Z }
+status: stable
+---
+
+# Capability Live Codex Matrix
+
+- Provider model: openai/gpt-5.5
+- Groups: 16
+- Cases: ap:ap-approval-deny-01, ar:ar-no-done-without-upload-01, bl:bl-cancelled-not-resolved-01, cm:cm-mention-structured-01, co:co-409-stop-01, dp:dp-base-revision-01, er:er-blocked-dedup-01, hb:hb-context-01, ix:ix-checkbox-01, mh:mh-blocked-handoff-01, rf:rf-api-404-report-01, rs:rs-dependency-blocked-wake-01, se:se-get-issue-01, st:st-backlog-park-01, su:su-crossteam-billing-01, wk:wk-ask-mode-01
+- Each row asserts expected and observed typed calls plus final mock state.
+- Scored report: `live-eval-slice-report.md`
diff --git a/packages/paperclip-runner/knowledge/evidence/capability/live-eval-slice-report.json b/packages/paperclip-runner/knowledge/evidence/capability/live-eval-slice-report.json
new file mode 100644
index 00000000000..a13b3a43d3b
--- /dev/null
+++ b/packages/paperclip-runner/knowledge/evidence/capability/live-eval-slice-report.json
@@ -0,0 +1,1544 @@
+{
+  "schema": "paperclip.runner.eval-slice-report.v1",
+  "bundle": {
+    "id": "evb-ccfa51b713b10e13",
+    "summary": "runnerd/codex-app-server@codex-app-server-v2 · gpt-5.5 · mock · 4 grants · prompt:capability-live-exact-call-v1 · faults:clean",
+    "declaration": {
+      "schema": "paperclip.runner.eval-bundle.v1",
+      "provider": {
+        "runtime": "runnerd",
+        "transport": "codex-app-server",
+        "protocolVersion": "codex-app-server-v2"
+      },
+      "model": {
+        "id": "gpt-5.5"
+      },
+      "launchContext": {
+        "workingDirectoryClass": "workspace-checkout",
+        "scenarioId": "capability-live-codex-matrix-v1",
+        "turnTimeoutMs": 60000
+      },
+      "promptPolicy": {
+        "id": "capability-live-exact-call-v1",
+        "callTemplate": "Call the named semantic operation exactly once with the declared JSON input.",
+        "restraintTemplate": "Do not call any tool for a control-plane-owned behavior."
+      },
+      "grants": [
+        "delegation:tasks:create",
+        "discovery:agents:read",
+        "discovery:tasks:read",
+        "governance:approvals:request"
+      ],
+      "runner": {
+        "package": "@paperclipai/paperclip-runner",
+        "binary": "paperclip-runnerd",
+        "version": "0.1.2"
+      },
+      "controlPlaneAdapter": {
+        "kind": "mock",
+        "contract": "paperclip.capability.mock-state.v1"
+      },
+      "faultInjection": []
+    }
+  },
+  "generatedFromLiveModel": true,
+  "cases": [
+    {
+      "observation": {
+        "caseId": "ap-approval-deny-01",
+        "provenance": {
+          "source": "live_model"
+        },
+        "controlPlaneOwned": false,
+        "expectedCalls": [
+          "request_approval"
+        ],
+        "observedCalls": [
+          "request_approval"
+        ],
+        "forbiddenCalls": [],
+        "finalState": {
+          "expected": "mutated",
+          "observed": "mutated"
+        },
+        "authorization": {
+          "expected": "allowed",
+          "observed": "allowed"
+        },
+        "trace": {
+          "runId": "run-capability-live-ap-approval-deny-01",
+          "sessionId": "session-capability-live-ap-approval-deny-01",
+          "turnId": "019ff1e7-5e9e-7353-abc9-734f85ed535b",
+          "itemId": "evidence-000022",
+          "receiptIds": [
+            "evidence-000023"
+          ],
+          "terminalPresent": true
+        },
+        "efficiency": {
+          "latencyMs": 4056,
+          "attempts": 1
+        },
+        "budget": {
+          "maxLatencyMs": 60000,
+          "maxAttempts": 1
+        }
+      },
+      "scorecard": {
+        "schema": "paperclip.runner.eval-scorecard.v1",
+        "bundleId": "evb-ccfa51b713b10e13",
+        "caseId": "ap-approval-deny-01",
+        "dimensions": {
+          "hard_invariants": {
+            "dimension": "hard_invariants",
+            "score": 1,
+            "passed": true,
+            "gate": true,
+            "weight": 0,
+            "reasons": []
+          },
+          "semantic_outcome": {
+            "dimension": "semantic_outcome",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.35,
+            "reasons": []
+          },
+          "trajectory_restraint": {
+            "dimension": "trajectory_restraint",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.3,
+            "reasons": []
+          },
+          "trace_completeness": {
+            "dimension": "trace_completeness",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.2,
+            "reasons": []
+          },
+          "quality_efficiency": {
+            "dimension": "quality_efficiency",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.15,
+            "reasons": []
+          }
+        },
+        "overall": {
+          "score": 1,
+          "gatePassed": true,
+          "passed": true
+        }
+      }
+    },
+    {
+      "observation": {
+        "caseId": "ar-no-done-without-upload-01",
+        "provenance": {
+          "source": "live_model"
+        },
+        "controlPlaneOwned": false,
+        "expectedCalls": [
+          "register_deliverable"
+        ],
+        "observedCalls": [
+          "register_deliverable"
+        ],
+        "forbiddenCalls": [],
+        "finalState": {
+          "expected": "mutated",
+          "observed": "mutated"
+        },
+        "authorization": {
+          "expected": "allowed",
+          "observed": "allowed"
+        },
+        "trace": {
+          "runId": "run-capability-live-ar-no-done-without-upload-01",
+          "sessionId": "session-capability-live-ar-no-done-without-upload-01",
+          "turnId": "019ff1e7-6ffb-7123-9f0a-37ca2f894aa2",
+          "itemId": "evidence-000022",
+          "receiptIds": [
+            "evidence-000023"
+          ],
+          "terminalPresent": true
+        },
+        "efficiency": {
+          "latencyMs": 3568,
+          "attempts": 1
+        },
+        "budget": {
+          "maxLatencyMs": 60000,
+          "maxAttempts": 1
+        }
+      },
+      "scorecard": {
+        "schema": "paperclip.runner.eval-scorecard.v1",
+        "bundleId": "evb-ccfa51b713b10e13",
+        "caseId": "ar-no-done-without-upload-01",
+        "dimensions": {
+          "hard_invariants": {
+            "dimension": "hard_invariants",
+            "score": 1,
+            "passed": true,
+            "gate": true,
+            "weight": 0,
+            "reasons": []
+          },
+          "semantic_outcome": {
+            "dimension": "semantic_outcome",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.35,
+            "reasons": []
+          },
+          "trajectory_restraint": {
+            "dimension": "trajectory_restraint",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.3,
+            "reasons": []
+          },
+          "trace_completeness": {
+            "dimension": "trace_completeness",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.2,
+            "reasons": []
+          },
+          "quality_efficiency": {
+            "dimension": "quality_efficiency",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.15,
+            "reasons": []
+          }
+        },
+        "overall": {
+          "score": 1,
+          "gatePassed": true,
+          "passed": true
+        }
+      }
+    },
+    {
+      "observation": {
+        "caseId": "bl-cancelled-not-resolved-01",
+        "provenance": {
+          "source": "live_model"
+        },
+        "controlPlaneOwned": false,
+        "expectedCalls": [
+          "block_task"
+        ],
+        "observedCalls": [
+          "block_task"
+        ],
+        "forbiddenCalls": [],
+        "finalState": {
+          "expected": "mutated",
+          "observed": "mutated"
+        },
+        "authorization": {
+          "expected": "allowed",
+          "observed": "allowed"
+        },
+        "trace": {
+          "runId": "run-capability-live-bl-cancelled-not-resolved-01",
+          "sessionId": "session-capability-live-bl-cancelled-not-resolved-01",
+          "turnId": "019ff1e7-7f46-7292-b3ab-5b5d6b99e09f",
+          "itemId": "evidence-000022",
+          "receiptIds": [
+            "evidence-000023"
+          ],
+          "terminalPresent": true
+        },
+        "efficiency": {
+          "latencyMs": 4079,
+          "attempts": 1
+        },
+        "budget": {
+          "maxLatencyMs": 60000,
+          "maxAttempts": 1
+        }
+      },
+      "scorecard": {
+        "schema": "paperclip.runner.eval-scorecard.v1",
+        "bundleId": "evb-ccfa51b713b10e13",
+        "caseId": "bl-cancelled-not-resolved-01",
+        "dimensions": {
+          "hard_invariants": {
+            "dimension": "hard_invariants",
+            "score": 1,
+            "passed": true,
+            "gate": true,
+            "weight": 0,
+            "reasons": []
+          },
+          "semantic_outcome": {
+            "dimension": "semantic_outcome",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.35,
+            "reasons": []
+          },
+          "trajectory_restraint": {
+            "dimension": "trajectory_restraint",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.3,
+            "reasons": []
+          },
+          "trace_completeness": {
+            "dimension": "trace_completeness",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.2,
+            "reasons": []
+          },
+          "quality_efficiency": {
+            "dimension": "quality_efficiency",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.15,
+            "reasons": []
+          }
+        },
+        "overall": {
+          "score": 1,
+          "gatePassed": true,
+          "passed": true
+        }
+      }
+    },
+    {
+      "observation": {
+        "caseId": "cm-mention-structured-01",
+        "provenance": {
+          "source": "live_model"
+        },
+        "controlPlaneOwned": false,
+        "expectedCalls": [
+          "report_progress"
+        ],
+        "observedCalls": [
+          "report_progress"
+        ],
+        "forbiddenCalls": [],
+        "finalState": {
+          "expected": "mutated",
+          "observed": "mutated"
+        },
+        "authorization": {
+          "expected": "allowed",
+          "observed": "allowed"
+        },
+        "trace": {
+          "runId": "run-capability-live-cm-mention-structured-01",
+          "sessionId": "session-capability-live-cm-mention-structured-01",
+          "turnId": "019ff1e7-909c-70e2-b9b7-8e53cd6dac67",
+          "itemId": "evidence-000021",
+          "receiptIds": [
+            "evidence-000023"
+          ],
+          "terminalPresent": true
+        },
+        "efficiency": {
+          "latencyMs": 3621,
+          "attempts": 1
+        },
+        "budget": {
+          "maxLatencyMs": 60000,
+          "maxAttempts": 1
+        }
+      },
+      "scorecard": {
+        "schema": "paperclip.runner.eval-scorecard.v1",
+        "bundleId": "evb-ccfa51b713b10e13",
+        "caseId": "cm-mention-structured-01",
+        "dimensions": {
+          "hard_invariants": {
+            "dimension": "hard_invariants",
+            "score": 1,
+            "passed": true,
+            "gate": true,
+            "weight": 0,
+            "reasons": []
+          },
+          "semantic_outcome": {
+            "dimension": "semantic_outcome",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.35,
+            "reasons": []
+          },
+          "trajectory_restraint": {
+            "dimension": "trajectory_restraint",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.3,
+            "reasons": []
+          },
+          "trace_completeness": {
+            "dimension": "trace_completeness",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.2,
+            "reasons": []
+          },
+          "quality_efficiency": {
+            "dimension": "quality_efficiency",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.15,
+            "reasons": []
+          }
+        },
+        "overall": {
+          "score": 1,
+          "gatePassed": true,
+          "passed": true
+        }
+      }
+    },
+    {
+      "observation": {
+        "caseId": "co-409-stop-01",
+        "provenance": {
+          "source": "live_model"
+        },
+        "controlPlaneOwned": true,
+        "expectedCalls": [],
+        "observedCalls": [],
+        "forbiddenCalls": [],
+        "finalState": {
+          "expected": "unchanged",
+          "observed": "unchanged"
+        },
+        "authorization": {
+          "expected": "absent",
+          "observed": "absent"
+        },
+        "trace": {
+          "runId": "run-capability-live-co-409-stop-01",
+          "sessionId": "session-capability-live-co-409-stop-01",
+          "turnId": "019ff1e7-a051-7760-992c-39e9c9f2b09c",
+          "itemId": "transcript-000023",
+          "receiptIds": [],
+          "terminalPresent": true
+        },
+        "efficiency": {
+          "latencyMs": 1864,
+          "attempts": 1
+        },
+        "budget": {
+          "maxLatencyMs": 60000,
+          "maxAttempts": 1
+        }
+      },
+      "scorecard": {
+        "schema": "paperclip.runner.eval-scorecard.v1",
+        "bundleId": "evb-ccfa51b713b10e13",
+        "caseId": "co-409-stop-01",
+        "dimensions": {
+          "hard_invariants": {
+            "dimension": "hard_invariants",
+            "score": 1,
+            "passed": true,
+            "gate": true,
+            "weight": 0,
+            "reasons": []
+          },
+          "semantic_outcome": {
+            "dimension": "semantic_outcome",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.35,
+            "reasons": []
+          },
+          "trajectory_restraint": {
+            "dimension": "trajectory_restraint",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.3,
+            "reasons": [
+              "restraint honored: no call made"
+            ]
+          },
+          "trace_completeness": {
+            "dimension": "trace_completeness",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.2,
+            "reasons": []
+          },
+          "quality_efficiency": {
+            "dimension": "quality_efficiency",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.15,
+            "reasons": []
+          }
+        },
+        "overall": {
+          "score": 1,
+          "gatePassed": true,
+          "passed": true
+        }
+      }
+    },
+    {
+      "observation": {
+        "caseId": "dp-base-revision-01",
+        "provenance": {
+          "source": "live_model"
+        },
+        "controlPlaneOwned": false,
+        "expectedCalls": [
+          "write_document"
+        ],
+        "observedCalls": [
+          "write_document"
+        ],
+        "forbiddenCalls": [],
+        "finalState": {
+          "expected": "mutated",
+          "observed": "mutated"
+        },
+        "authorization": {
+          "expected": "allowed",
+          "observed": "allowed"
+        },
+        "trace": {
+          "runId": "run-capability-live-dp-base-revision-01",
+          "sessionId": "session-capability-live-dp-base-revision-01",
+          "turnId": "019ff1e7-a905-7551-b66a-08d1fd0832fd",
+          "itemId": "evidence-000022",
+          "receiptIds": [
+            "evidence-000023"
+          ],
+          "terminalPresent": true
+        },
+        "efficiency": {
+          "latencyMs": 3685,
+          "attempts": 1
+        },
+        "budget": {
+          "maxLatencyMs": 60000,
+          "maxAttempts": 1
+        }
+      },
+      "scorecard": {
+        "schema": "paperclip.runner.eval-scorecard.v1",
+        "bundleId": "evb-ccfa51b713b10e13",
+        "caseId": "dp-base-revision-01",
+        "dimensions": {
+          "hard_invariants": {
+            "dimension": "hard_invariants",
+            "score": 1,
+            "passed": true,
+            "gate": true,
+            "weight": 0,
+            "reasons": []
+          },
+          "semantic_outcome": {
+            "dimension": "semantic_outcome",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.35,
+            "reasons": []
+          },
+          "trajectory_restraint": {
+            "dimension": "trajectory_restraint",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.3,
+            "reasons": []
+          },
+          "trace_completeness": {
+            "dimension": "trace_completeness",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.2,
+            "reasons": []
+          },
+          "quality_efficiency": {
+            "dimension": "quality_efficiency",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.15,
+            "reasons": []
+          }
+        },
+        "overall": {
+          "score": 1,
+          "gatePassed": true,
+          "passed": true
+        }
+      }
+    },
+    {
+      "observation": {
+        "caseId": "er-blocked-dedup-01",
+        "provenance": {
+          "source": "live_model"
+        },
+        "controlPlaneOwned": true,
+        "expectedCalls": [],
+        "observedCalls": [],
+        "forbiddenCalls": [],
+        "finalState": {
+          "expected": "unchanged",
+          "observed": "unchanged"
+        },
+        "authorization": {
+          "expected": "absent",
+          "observed": "absent"
+        },
+        "trace": {
+          "runId": "run-capability-live-er-blocked-dedup-01",
+          "sessionId": "session-capability-live-er-blocked-dedup-01",
+          "turnId": "019ff1e7-b8d6-7eb2-9d4d-e98b659fdfc0",
+          "itemId": "transcript-000023",
+          "receiptIds": [],
+          "terminalPresent": true
+        },
+        "efficiency": {
+          "latencyMs": 2033,
+          "attempts": 1
+        },
+        "budget": {
+          "maxLatencyMs": 60000,
+          "maxAttempts": 1
+        }
+      },
+      "scorecard": {
+        "schema": "paperclip.runner.eval-scorecard.v1",
+        "bundleId": "evb-ccfa51b713b10e13",
+        "caseId": "er-blocked-dedup-01",
+        "dimensions": {
+          "hard_invariants": {
+            "dimension": "hard_invariants",
+            "score": 1,
+            "passed": true,
+            "gate": true,
+            "weight": 0,
+            "reasons": []
+          },
+          "semantic_outcome": {
+            "dimension": "semantic_outcome",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.35,
+            "reasons": []
+          },
+          "trajectory_restraint": {
+            "dimension": "trajectory_restraint",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.3,
+            "reasons": [
+              "restraint honored: no call made"
+            ]
+          },
+          "trace_completeness": {
+            "dimension": "trace_completeness",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.2,
+            "reasons": []
+          },
+          "quality_efficiency": {
+            "dimension": "quality_efficiency",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.15,
+            "reasons": []
+          }
+        },
+        "overall": {
+          "score": 1,
+          "gatePassed": true,
+          "passed": true
+        }
+      }
+    },
+    {
+      "observation": {
+        "caseId": "hb-context-01",
+        "provenance": {
+          "source": "live_model"
+        },
+        "controlPlaneOwned": true,
+        "expectedCalls": [],
+        "observedCalls": [],
+        "forbiddenCalls": [],
+        "finalState": {
+          "expected": "unchanged",
+          "observed": "unchanged"
+        },
+        "authorization": {
+          "expected": "absent",
+          "observed": "absent"
+        },
+        "trace": {
+          "runId": "run-capability-live-hb-context-01",
+          "sessionId": "session-capability-live-hb-context-01",
+          "turnId": "019ff1e7-c239-7462-8d23-dda298d0fa58",
+          "itemId": "transcript-000023",
+          "receiptIds": [],
+          "terminalPresent": true
+        },
+        "efficiency": {
+          "latencyMs": 1658,
+          "attempts": 1
+        },
+        "budget": {
+          "maxLatencyMs": 60000,
+          "maxAttempts": 1
+        }
+      },
+      "scorecard": {
+        "schema": "paperclip.runner.eval-scorecard.v1",
+        "bundleId": "evb-ccfa51b713b10e13",
+        "caseId": "hb-context-01",
+        "dimensions": {
+          "hard_invariants": {
+            "dimension": "hard_invariants",
+            "score": 1,
+            "passed": true,
+            "gate": true,
+            "weight": 0,
+            "reasons": []
+          },
+          "semantic_outcome": {
+            "dimension": "semantic_outcome",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.35,
+            "reasons": []
+          },
+          "trajectory_restraint": {
+            "dimension": "trajectory_restraint",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.3,
+            "reasons": [
+              "restraint honored: no call made"
+            ]
+          },
+          "trace_completeness": {
+            "dimension": "trace_completeness",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.2,
+            "reasons": []
+          },
+          "quality_efficiency": {
+            "dimension": "quality_efficiency",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.15,
+            "reasons": []
+          }
+        },
+        "overall": {
+          "score": 1,
+          "gatePassed": true,
+          "passed": true
+        }
+      }
+    },
+    {
+      "observation": {
+        "caseId": "ix-checkbox-01",
+        "provenance": {
+          "source": "live_model"
+        },
+        "controlPlaneOwned": false,
+        "expectedCalls": [
+          "request_human_input"
+        ],
+        "observedCalls": [
+          "request_human_input"
+        ],
+        "forbiddenCalls": [],
+        "finalState": {
+          "expected": "mutated",
+          "observed": "mutated"
+        },
+        "authorization": {
+          "expected": "allowed",
+          "observed": "allowed"
+        },
+        "trace": {
+          "runId": "run-capability-live-ix-checkbox-01",
+          "sessionId": "session-capability-live-ix-checkbox-01",
+          "turnId": "019ff1e7-ca3a-7ad2-9d67-35ee17cb6042",
+          "itemId": "evidence-000021",
+          "receiptIds": [
+            "evidence-000023"
+          ],
+          "terminalPresent": true
+        },
+        "efficiency": {
+          "latencyMs": 9631,
+          "attempts": 1
+        },
+        "budget": {
+          "maxLatencyMs": 60000,
+          "maxAttempts": 1
+        }
+      },
+      "scorecard": {
+        "schema": "paperclip.runner.eval-scorecard.v1",
+        "bundleId": "evb-ccfa51b713b10e13",
+        "caseId": "ix-checkbox-01",
+        "dimensions": {
+          "hard_invariants": {
+            "dimension": "hard_invariants",
+            "score": 1,
+            "passed": true,
+            "gate": true,
+            "weight": 0,
+            "reasons": []
+          },
+          "semantic_outcome": {
+            "dimension": "semantic_outcome",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.35,
+            "reasons": []
+          },
+          "trajectory_restraint": {
+            "dimension": "trajectory_restraint",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.3,
+            "reasons": []
+          },
+          "trace_completeness": {
+            "dimension": "trace_completeness",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.2,
+            "reasons": []
+          },
+          "quality_efficiency": {
+            "dimension": "quality_efficiency",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.15,
+            "reasons": []
+          }
+        },
+        "overall": {
+          "score": 1,
+          "gatePassed": true,
+          "passed": true
+        }
+      }
+    },
+    {
+      "observation": {
+        "caseId": "mh-blocked-handoff-01",
+        "provenance": {
+          "source": "live_model"
+        },
+        "controlPlaneOwned": false,
+        "expectedCalls": [
+          "create_task"
+        ],
+        "observedCalls": [
+          "create_task"
+        ],
+        "forbiddenCalls": [],
+        "finalState": {
+          "expected": "mutated",
+          "observed": "mutated"
+        },
+        "authorization": {
+          "expected": "allowed",
+          "observed": "allowed"
+        },
+        "trace": {
+          "runId": "run-capability-live-mh-blocked-handoff-01",
+          "sessionId": "session-capability-live-mh-blocked-handoff-01",
+          "turnId": "019ff1e7-f14e-7962-91e0-60f4b2c1ee5e",
+          "itemId": "evidence-000022",
+          "receiptIds": [
+            "evidence-000023"
+          ],
+          "terminalPresent": true
+        },
+        "efficiency": {
+          "latencyMs": 3474,
+          "attempts": 1
+        },
+        "budget": {
+          "maxLatencyMs": 60000,
+          "maxAttempts": 1
+        }
+      },
+      "scorecard": {
+        "schema": "paperclip.runner.eval-scorecard.v1",
+        "bundleId": "evb-ccfa51b713b10e13",
+        "caseId": "mh-blocked-handoff-01",
+        "dimensions": {
+          "hard_invariants": {
+            "dimension": "hard_invariants",
+            "score": 1,
+            "passed": true,
+            "gate": true,
+            "weight": 0,
+            "reasons": []
+          },
+          "semantic_outcome": {
+            "dimension": "semantic_outcome",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.35,
+            "reasons": []
+          },
+          "trajectory_restraint": {
+            "dimension": "trajectory_restraint",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.3,
+            "reasons": []
+          },
+          "trace_completeness": {
+            "dimension": "trace_completeness",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.2,
+            "reasons": []
+          },
+          "quality_efficiency": {
+            "dimension": "quality_efficiency",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.15,
+            "reasons": []
+          }
+        },
+        "overall": {
+          "score": 1,
+          "gatePassed": true,
+          "passed": true
+        }
+      }
+    },
+    {
+      "observation": {
+        "caseId": "rf-api-404-report-01",
+        "provenance": {
+          "source": "live_model"
+        },
+        "controlPlaneOwned": false,
+        "expectedCalls": [
+          "list_agents"
+        ],
+        "observedCalls": [
+          "list_agents"
+        ],
+        "forbiddenCalls": [],
+        "finalState": {
+          "expected": "unchanged",
+          "observed": "unchanged"
+        },
+        "authorization": {
+          "expected": "allowed",
+          "observed": "allowed"
+        },
+        "trace": {
+          "runId": "run-capability-live-rf-api-404-report-01",
+          "sessionId": "session-capability-live-rf-api-404-report-01",
+          "turnId": "019ff1e8-0054-7102-8629-5a629712bc29",
+          "itemId": "evidence-000023",
+          "receiptIds": [
+            "evidence-000025"
+          ],
+          "terminalPresent": true
+        },
+        "efficiency": {
+          "latencyMs": 5691,
+          "attempts": 1
+        },
+        "budget": {
+          "maxLatencyMs": 60000,
+          "maxAttempts": 1
+        }
+      },
+      "scorecard": {
+        "schema": "paperclip.runner.eval-scorecard.v1",
+        "bundleId": "evb-ccfa51b713b10e13",
+        "caseId": "rf-api-404-report-01",
+        "dimensions": {
+          "hard_invariants": {
+            "dimension": "hard_invariants",
+            "score": 1,
+            "passed": true,
+            "gate": true,
+            "weight": 0,
+            "reasons": []
+          },
+          "semantic_outcome": {
+            "dimension": "semantic_outcome",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.35,
+            "reasons": []
+          },
+          "trajectory_restraint": {
+            "dimension": "trajectory_restraint",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.3,
+            "reasons": []
+          },
+          "trace_completeness": {
+            "dimension": "trace_completeness",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.2,
+            "reasons": []
+          },
+          "quality_efficiency": {
+            "dimension": "quality_efficiency",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.15,
+            "reasons": []
+          }
+        },
+        "overall": {
+          "score": 1,
+          "gatePassed": true,
+          "passed": true
+        }
+      }
+    },
+    {
+      "observation": {
+        "caseId": "rs-dependency-blocked-wake-01",
+        "provenance": {
+          "source": "live_model"
+        },
+        "controlPlaneOwned": true,
+        "expectedCalls": [],
+        "observedCalls": [],
+        "forbiddenCalls": [],
+        "finalState": {
+          "expected": "unchanged",
+          "observed": "unchanged"
+        },
+        "authorization": {
+          "expected": "absent",
+          "observed": "absent"
+        },
+        "trace": {
+          "runId": "run-capability-live-rs-dependency-blocked-wake-01",
+          "sessionId": "session-capability-live-rs-dependency-blocked-wake-01",
+          "turnId": "019ff1e8-17fc-7600-aa46-84429760dcdd",
+          "itemId": "transcript-000023",
+          "receiptIds": [],
+          "terminalPresent": true
+        },
+        "efficiency": {
+          "latencyMs": 1908,
+          "attempts": 1
+        },
+        "budget": {
+          "maxLatencyMs": 60000,
+          "maxAttempts": 1
+        }
+      },
+      "scorecard": {
+        "schema": "paperclip.runner.eval-scorecard.v1",
+        "bundleId": "evb-ccfa51b713b10e13",
+        "caseId": "rs-dependency-blocked-wake-01",
+        "dimensions": {
+          "hard_invariants": {
+            "dimension": "hard_invariants",
+            "score": 1,
+            "passed": true,
+            "gate": true,
+            "weight": 0,
+            "reasons": []
+          },
+          "semantic_outcome": {
+            "dimension": "semantic_outcome",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.35,
+            "reasons": []
+          },
+          "trajectory_restraint": {
+            "dimension": "trajectory_restraint",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.3,
+            "reasons": [
+              "restraint honored: no call made"
+            ]
+          },
+          "trace_completeness": {
+            "dimension": "trace_completeness",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.2,
+            "reasons": []
+          },
+          "quality_efficiency": {
+            "dimension": "quality_efficiency",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.15,
+            "reasons": []
+          }
+        },
+        "overall": {
+          "score": 1,
+          "gatePassed": true,
+          "passed": true
+        }
+      }
+    },
+    {
+      "observation": {
+        "caseId": "se-get-issue-01",
+        "provenance": {
+          "source": "live_model"
+        },
+        "controlPlaneOwned": false,
+        "expectedCalls": [
+          "search_tasks"
+        ],
+        "observedCalls": [
+          "search_tasks"
+        ],
+        "forbiddenCalls": [],
+        "finalState": {
+          "expected": "unchanged",
+          "observed": "unchanged"
+        },
+        "authorization": {
+          "expected": "allowed",
+          "observed": "denied"
+        },
+        "trace": {
+          "runId": "run-capability-live-se-get-issue-01",
+          "sessionId": "session-capability-live-se-get-issue-01",
+          "turnId": "019ff1e8-20ea-73e2-9cc6-3e2d8f6fc948",
+          "itemId": "evidence-000024",
+          "receiptIds": [
+            "evidence-000025"
+          ],
+          "terminalPresent": true
+        },
+        "efficiency": {
+          "latencyMs": 4504,
+          "attempts": 1
+        },
+        "budget": {
+          "maxLatencyMs": 60000,
+          "maxAttempts": 1
+        }
+      },
+      "scorecard": {
+        "schema": "paperclip.runner.eval-scorecard.v1",
+        "bundleId": "evb-ccfa51b713b10e13",
+        "caseId": "se-get-issue-01",
+        "dimensions": {
+          "hard_invariants": {
+            "dimension": "hard_invariants",
+            "score": 1,
+            "passed": true,
+            "gate": true,
+            "weight": 0,
+            "reasons": []
+          },
+          "semantic_outcome": {
+            "dimension": "semantic_outcome",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.35,
+            "reasons": []
+          },
+          "trajectory_restraint": {
+            "dimension": "trajectory_restraint",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.3,
+            "reasons": []
+          },
+          "trace_completeness": {
+            "dimension": "trace_completeness",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.2,
+            "reasons": []
+          },
+          "quality_efficiency": {
+            "dimension": "quality_efficiency",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.15,
+            "reasons": []
+          }
+        },
+        "overall": {
+          "score": 1,
+          "gatePassed": true,
+          "passed": true
+        }
+      }
+    },
+    {
+      "observation": {
+        "caseId": "st-backlog-park-01",
+        "provenance": {
+          "source": "live_model"
+        },
+        "controlPlaneOwned": false,
+        "expectedCalls": [
+          "finish_task"
+        ],
+        "observedCalls": [
+          "finish_task"
+        ],
+        "forbiddenCalls": [],
+        "finalState": {
+          "expected": "mutated",
+          "observed": "mutated"
+        },
+        "authorization": {
+          "expected": "allowed",
+          "observed": "allowed"
+        },
+        "trace": {
+          "runId": "run-capability-live-st-backlog-park-01",
+          "sessionId": "session-capability-live-st-backlog-park-01",
+          "turnId": "019ff1e8-33f6-7f93-88b9-83ad36eac2bb",
+          "itemId": "evidence-000022",
+          "receiptIds": [
+            "evidence-000023"
+          ],
+          "terminalPresent": true
+        },
+        "efficiency": {
+          "latencyMs": 4513,
+          "attempts": 1
+        },
+        "budget": {
+          "maxLatencyMs": 60000,
+          "maxAttempts": 1
+        }
+      },
+      "scorecard": {
+        "schema": "paperclip.runner.eval-scorecard.v1",
+        "bundleId": "evb-ccfa51b713b10e13",
+        "caseId": "st-backlog-park-01",
+        "dimensions": {
+          "hard_invariants": {
+            "dimension": "hard_invariants",
+            "score": 1,
+            "passed": true,
+            "gate": true,
+            "weight": 0,
+            "reasons": []
+          },
+          "semantic_outcome": {
+            "dimension": "semantic_outcome",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.35,
+            "reasons": []
+          },
+          "trajectory_restraint": {
+            "dimension": "trajectory_restraint",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.3,
+            "reasons": []
+          },
+          "trace_completeness": {
+            "dimension": "trace_completeness",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.2,
+            "reasons": []
+          },
+          "quality_efficiency": {
+            "dimension": "quality_efficiency",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.15,
+            "reasons": []
+          }
+        },
+        "overall": {
+          "score": 1,
+          "gatePassed": true,
+          "passed": true
+        }
+      }
+    },
+    {
+      "observation": {
+        "caseId": "su-crossteam-billing-01",
+        "provenance": {
+          "source": "live_model"
+        },
+        "controlPlaneOwned": false,
+        "expectedCalls": [
+          "create_task"
+        ],
+        "observedCalls": [
+          "create_task"
+        ],
+        "forbiddenCalls": [],
+        "finalState": {
+          "expected": "mutated",
+          "observed": "mutated"
+        },
+        "authorization": {
+          "expected": "allowed",
+          "observed": "allowed"
+        },
+        "trace": {
+          "runId": "run-capability-live-su-crossteam-billing-01",
+          "sessionId": "session-capability-live-su-crossteam-billing-01",
+          "turnId": "019ff1e8-470b-7550-a6b6-8c0c842eb948",
+          "itemId": "evidence-000022",
+          "receiptIds": [
+            "evidence-000023"
+          ],
+          "terminalPresent": true
+        },
+        "efficiency": {
+          "latencyMs": 6353,
+          "attempts": 1
+        },
+        "budget": {
+          "maxLatencyMs": 60000,
+          "maxAttempts": 1
+        }
+      },
+      "scorecard": {
+        "schema": "paperclip.runner.eval-scorecard.v1",
+        "bundleId": "evb-ccfa51b713b10e13",
+        "caseId": "su-crossteam-billing-01",
+        "dimensions": {
+          "hard_invariants": {
+            "dimension": "hard_invariants",
+            "score": 1,
+            "passed": true,
+            "gate": true,
+            "weight": 0,
+            "reasons": []
+          },
+          "semantic_outcome": {
+            "dimension": "semantic_outcome",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.35,
+            "reasons": []
+          },
+          "trajectory_restraint": {
+            "dimension": "trajectory_restraint",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.3,
+            "reasons": []
+          },
+          "trace_completeness": {
+            "dimension": "trace_completeness",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.2,
+            "reasons": []
+          },
+          "quality_efficiency": {
+            "dimension": "quality_efficiency",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.15,
+            "reasons": []
+          }
+        },
+        "overall": {
+          "score": 1,
+          "gatePassed": true,
+          "passed": true
+        }
+      }
+    },
+    {
+      "observation": {
+        "caseId": "wk-ask-mode-01",
+        "provenance": {
+          "source": "live_model"
+        },
+        "controlPlaneOwned": true,
+        "expectedCalls": [],
+        "observedCalls": [],
+        "forbiddenCalls": [],
+        "finalState": {
+          "expected": "unchanged",
+          "observed": "unchanged"
+        },
+        "authorization": {
+          "expected": "absent",
+          "observed": "absent"
+        },
+        "trace": {
+          "runId": "run-capability-live-wk-ask-mode-01",
+          "sessionId": "session-capability-live-wk-ask-mode-01",
+          "turnId": "019ff1e8-617d-7ba3-8c12-84e6c5d011d0",
+          "itemId": "transcript-000023",
+          "receiptIds": [],
+          "terminalPresent": true
+        },
+        "efficiency": {
+          "latencyMs": 1731,
+          "attempts": 1
+        },
+        "budget": {
+          "maxLatencyMs": 60000,
+          "maxAttempts": 1
+        }
+      },
+      "scorecard": {
+        "schema": "paperclip.runner.eval-scorecard.v1",
+        "bundleId": "evb-ccfa51b713b10e13",
+        "caseId": "wk-ask-mode-01",
+        "dimensions": {
+          "hard_invariants": {
+            "dimension": "hard_invariants",
+            "score": 1,
+            "passed": true,
+            "gate": true,
+            "weight": 0,
+            "reasons": []
+          },
+          "semantic_outcome": {
+            "dimension": "semantic_outcome",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.35,
+            "reasons": []
+          },
+          "trajectory_restraint": {
+            "dimension": "trajectory_restraint",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.3,
+            "reasons": [
+              "restraint honored: no call made"
+            ]
+          },
+          "trace_completeness": {
+            "dimension": "trace_completeness",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.2,
+            "reasons": []
+          },
+          "quality_efficiency": {
+            "dimension": "quality_efficiency",
+            "score": 1,
+            "passed": true,
+            "gate": false,
+            "weight": 0.15,
+            "reasons": []
+          }
+        },
+        "overall": {
+          "score": 1,
+          "gatePassed": true,
+          "passed": true
+        }
+      }
+    }
+  ],
+  "aggregate": {
+    "caseCount": 16,
+    "passed": 16,
+    "gateFailures": 0,
+    "meanOverall": 1,
+    "dimensionMeans": {
+      "hard_invariants": 1,
+      "semantic_outcome": 1,
+      "trajectory_restraint": 1,
+      "trace_completeness": 1,
+      "quality_efficiency": 1
+    }
+  }
+}
diff --git a/packages/paperclip-runner/knowledge/evidence/capability/live-eval-slice-report.md b/packages/paperclip-runner/knowledge/evidence/capability/live-eval-slice-report.md
new file mode 100644
index 00000000000..448afaba40e
--- /dev/null
+++ b/packages/paperclip-runner/knowledge/evidence/capability/live-eval-slice-report.md
@@ -0,0 +1,42 @@
+---
+type: Evidence Report
+title: Live Runner Eval Slice Report
+description: Scored real-Codex capability matrix bound to its secret-free eval bundle.
+generated: { by: process:capability-live-eval, at: 2026-08-11T17:38:30.000Z }
+status: stable
+---
+
+# Runner eval slice — evb-ccfa51b713b10e13
+
+- Bundle: `runnerd/codex-app-server@codex-app-server-v2 · gpt-5.5 · mock · 4 grants · prompt:capability-live-exact-call-v1 · faults:clean`
+- Source: real model tool choice
+- Cases: 16 · passed 16 · gate failures 0 · mean overall 1
+
+## Dimension means
+
+- hard_invariants: 1
+- semantic_outcome: 1
+- trajectory_restraint: 1
+- trace_completeness: 1
+- quality_efficiency: 1
+
+## Cases
+
+| case | source | counterpart | fault | gate | overall | outcome | trajectory | trace | efficiency |
+| --- | --- | --- | --- | --- | --- | --- | --- | --- | --- |
+| ap-approval-deny-01 | live_model | — | — | ok | 1 | 1 | 1 | 1 | 1 |
+| ar-no-done-without-upload-01 | live_model | — | — | ok | 1 | 1 | 1 | 1 | 1 |
+| bl-cancelled-not-resolved-01 | live_model | — | — | ok | 1 | 1 | 1 | 1 | 1 |
+| cm-mention-structured-01 | live_model | — | — | ok | 1 | 1 | 1 | 1 | 1 |
+| co-409-stop-01 | live_model | — | — | ok | 1 | 1 | 1 | 1 | 1 |
+| dp-base-revision-01 | live_model | — | — | ok | 1 | 1 | 1 | 1 | 1 |
+| er-blocked-dedup-01 | live_model | — | — | ok | 1 | 1 | 1 | 1 | 1 |
+| hb-context-01 | live_model | — | — | ok | 1 | 1 | 1 | 1 | 1 |
+| ix-checkbox-01 | live_model | — | — | ok | 1 | 1 | 1 | 1 | 1 |
+| mh-blocked-handoff-01 | live_model | — | — | ok | 1 | 1 | 1 | 1 | 1 |
+| rf-api-404-report-01 | live_model | — | — | ok | 1 | 1 | 1 | 1 | 1 |
+| rs-dependency-blocked-wake-01 | live_model | — | — | ok | 1 | 1 | 1 | 1 | 1 |
+| se-get-issue-01 | live_model | — | — | ok | 1 | 1 | 1 | 1 | 1 |
+| st-backlog-park-01 | live_model | — | — | ok | 1 | 1 | 1 | 1 | 1 |
+| su-crossteam-billing-01 | live_model | — | — | ok | 1 | 1 | 1 | 1 | 1 |
+| wk-ask-mode-01 | live_model | — | — | ok | 1 | 1 | 1 | 1 | 1 |
diff --git a/packages/paperclip-runner/knowledge/evidence/index.md b/packages/paperclip-runner/knowledge/evidence/index.md
index 1d748214630..615fa96a7a8 100644
--- a/packages/paperclip-runner/knowledge/evidence/index.md
+++ b/packages/paperclip-runner/knowledge/evidence/index.md
@@ -8,6 +8,7 @@ records that remain useful to package verification and debugging.
 - [Durable recovery per-fault traces](durable-recovery-faults/)
 - [Codex normalized trace](codex-trace.json)
 - [Live-console real Codex server trace](live-console-real-codex-server.json)
+- [Capability live eval evidence](capability/)
 
 Screenshots, dated milestone reports, and duplicate handoff records live in Git
 history rather than the current package.
diff --git a/packages/paperclip-runner/scripts/run-capability-live-eval-matrix.mjs b/packages/paperclip-runner/scripts/run-capability-live-eval-matrix.mjs
index 6cfa41b873c..2391d9ebba6 100644
--- a/packages/paperclip-runner/scripts/run-capability-live-eval-matrix.mjs
+++ b/packages/paperclip-runner/scripts/run-capability-live-eval-matrix.mjs
@@ -1,27 +1,157 @@
-import { mkdir, writeFile } from "node:fs/promises";
+import { mkdir, readFile, writeFile } from "node:fs/promises";
 import { dirname, resolve } from "node:path";
 
 const packageRoot = resolve(import.meta.dirname, "..");
 const suite = await import(resolve(packageRoot, "dist/conformance/capability-eval-suite.js"));
+const evals = await import(resolve(packageRoot, "dist/eval/index.js"));
+const packageManifest = JSON.parse(await readFile(resolve(packageRoot, "package.json"), "utf8"));
+const generatedAt = new Date().toISOString();
+const evidenceMarkdown = (title, description, body) => [
+  "---",
+  "type: Evidence Report",
+  `title: ${title}`,
+  `description: ${description}`,
+  `generated: { by: process:capability-live-eval, at: ${generatedAt} }`,
+  "status: stable",
+  "---",
+  "",
+  body.trimEnd(),
+  "",
+].join("\n");
 const matrix = await suite.runCapabilityLiveCodexMatrix(process.cwd());
+const providerModels = [...new Set(matrix.map(
+  (entry) => `${entry.providerModel.provider}/${entry.providerModel.id}`,
+))];
+if (providerModels.length !== 1) {
+  throw new Error(`live eval matrix used inconsistent provider models: ${providerModels.join(", ")}`);
+}
+const liveModel = matrix[0].providerModel;
+const liveBundle = {
+  schema: evals.EVAL_BUNDLE_SCHEMA,
+  provider: {
+    runtime: "runnerd",
+    transport: "codex-app-server",
+    protocolVersion: "codex-app-server-v2",
+  },
+  model: { id: liveModel.id },
+  launchContext: {
+    workingDirectoryClass: "workspace-checkout",
+    scenarioId: "capability-live-codex-matrix-v1",
+    turnTimeoutMs: 60_000,
+  },
+  promptPolicy: {
+    id: "capability-live-exact-call-v1",
+    callTemplate: "Call the named semantic operation exactly once with the declared JSON input.",
+    restraintTemplate: "Do not call any tool for a control-plane-owned behavior.",
+  },
+  grants: [...suite.CAPABILITY_LIVE_EVAL_GRANTS],
+  runner: {
+    package: packageManifest.name,
+    binary: "paperclip-runnerd",
+    version: packageManifest.version,
+  },
+  controlPlaneAdapter: {
+    kind: "mock",
+    contract: "paperclip.capability.mock-state.v1",
+  },
+  faultInjection: [],
+};
+const liveObservations = matrix.map((result) => evals.observationFromLiveMatrix(result, {
+  ...result.scoringEvidence,
+  provenance: { source: "live_model" },
+}));
+const liveSliceReport = evals.buildEvalSliceReport(liveBundle, liveObservations, {
+  generatedFromLiveModel: true,
+});
+if (liveSliceReport.aggregate.passed !== liveSliceReport.aggregate.caseCount) {
+  throw new Error(`live eval slice scored ${liveSliceReport.aggregate.passed}/${liveSliceReport.aggregate.caseCount} passing cases`);
+}
+
+const faultBundle = {
+  schema: evals.EVAL_BUNDLE_SCHEMA,
+  provider: {
+    runtime: "deterministic-harness",
+    transport: "in-process",
+    protocolVersion: "v1",
+  },
+  model: { id: "scripted-eval-agent" },
+  launchContext: {
+    workingDirectoryClass: "ephemeral-fixture",
+    scenarioId: "eval-behavior-fault-matrix-v1",
+    turnTimeoutMs: 1_000,
+  },
+  promptPolicy: {
+    id: "deterministic-behavior-driver-v1",
+    callTemplate: "Invoke the declared semantic operation sequence.",
+    restraintTemplate: "Make no semantic call.",
+  },
+  grants: ["control_plane:wakes", "governance:approvals:decide"],
+  runner: {
+    package: packageManifest.name,
+    binary: "node",
+    version: packageManifest.version,
+  },
+  controlPlaneAdapter: {
+    kind: "mock",
+    contract: "paperclip.capability.mock-state.v1",
+  },
+  faultInjection: [
+    { id: "fault-authz", class: "authorization", description: "deny approval decision exposure" },
+    { id: "fault-conflict", class: "conflict", description: "submit a stale plan revision" },
+    { id: "fault-retry", class: "retry", description: "fail the first artifact registration attempt" },
+    { id: "fault-provider", class: "provider_capability", description: "omit wake scheduling from the tool surface" },
+  ],
+};
+const behaviorMatrix = await evals.runEvalBehaviorFaultMatrix(faultBundle);
+const faultSliceReport = evals.buildEvalSliceReport(
+  faultBundle,
+  behaviorMatrix.map((result) => result.observation),
+);
 const report = {
   schema: "paperclip.capability.live-codex-matrix.v1",
   groups: matrix.length,
+  providerModel: liveModel,
   matrix,
 };
-const output = resolve(packageRoot, "knowledge/evidence/capability/live-codex-matrix.json");
-const markdown = output.replace(/\.json$/, ".md");
+const evidenceDirectory = resolve(packageRoot, "knowledge/evidence/capability");
+const output = resolve(evidenceDirectory, "live-codex-matrix.json");
+const markdown = resolve(evidenceDirectory, "live-codex-matrix.md");
+const liveSliceJson = resolve(evidenceDirectory, "live-eval-slice-report.json");
+const liveSliceMarkdown = resolve(evidenceDirectory, "live-eval-slice-report.md");
+const faultSliceJson = resolve(evidenceDirectory, "fault-eval-slice-report.json");
+const faultSliceMarkdown = resolve(evidenceDirectory, "fault-eval-slice-report.md");
 await mkdir(dirname(output), { recursive: true });
 await Promise.all([
   writeFile(output, `${JSON.stringify(report, null, 2)}\n`),
-  writeFile(markdown, [
+  writeFile(markdown, evidenceMarkdown(
+    "Capability Live Codex Matrix",
+    "Real Codex semantic-tool choices and resulting mock control-plane state.",
+    [
     "# Capability Live Codex Matrix",
     "",
+    `- Provider model: ${liveModel.provider}/${liveModel.id}`,
     `- Groups: ${report.groups}`,
     `- Cases: ${matrix.map((entry) => `${entry.group}:${entry.caseId}`).join(", ")}`,
     "- Each row asserts expected and observed typed calls plus final mock state.",
+    "- Scored report: `live-eval-slice-report.md`",
     "",
-  ].join("\n")),
+    ].join("\n"),
+  )),
+  writeFile(liveSliceJson, `${JSON.stringify(liveSliceReport, null, 2)}\n`),
+  writeFile(liveSliceMarkdown, evidenceMarkdown(
+    "Live Runner Eval Slice Report",
+    "Scored real-Codex capability matrix bound to its secret-free eval bundle.",
+    evals.renderEvalSliceMarkdown(liveSliceReport),
+  )),
+  writeFile(faultSliceJson, `${JSON.stringify(faultSliceReport, null, 2)}\n`),
+  writeFile(faultSliceMarkdown, evidenceMarkdown(
+    "Runner Eval Fault Slice Report",
+    "Eight green/red behaviors with deterministic injected-fault receipts.",
+    evals.renderEvalSliceMarkdown(faultSliceReport),
+  )),
 ]);
 process.stdout.write(`Capability live Codex matrix passed: ${matrix.length} groups.\n`);
+process.stdout.write(`Live eval slice passed: ${liveSliceReport.aggregate.passed}/${liveSliceReport.aggregate.caseCount}.\n`);
+process.stdout.write(`Behavior fault matrix passed: ${faultSliceReport.aggregate.passed} green, ${faultSliceReport.aggregate.caseCount - faultSliceReport.aggregate.passed} scored red.\n`);
 process.stdout.write(`Live matrix reports: ${output} and ${markdown}\n`);
+process.stdout.write(`Scored eval reports: ${liveSliceJson}, ${liveSliceMarkdown}, ${faultSliceJson}, and ${faultSliceMarkdown}\n`);
diff --git a/packages/paperclip-runner/src/conformance/capability-eval-suite.ts b/packages/paperclip-runner/src/conformance/capability-eval-suite.ts
index e166529451e..2e1d6f7276f 100644
--- a/packages/paperclip-runner/src/conformance/capability-eval-suite.ts
+++ b/packages/paperclip-runner/src/conformance/capability-eval-suite.ts
@@ -58,8 +58,34 @@ export interface CapabilityLiveCodexMatrixResult {
   forbiddenCalls: string[];
   observedCalls: string[];
   finalState: { expected: "unchanged" | "mutated"; observed: "unchanged" | "mutated" };
+  providerModel: { id: string; provider: string };
+  /** Browser-safe evidence retained from the live-session snapshot for scoring. */
+  scoringEvidence: {
+    trace: {
+      runId: string;
+      sessionId: string;
+      turnId: string;
+      itemId: string;
+      receiptIds: string[];
+      terminalPresent: boolean;
+    };
+    authorization: {
+      expected: "allowed" | "denied" | "absent";
+      observed: "allowed" | "denied" | "absent";
+    };
+    efficiency: { latencyMs: number; attempts: number };
+    budget: { maxLatencyMs: number; maxAttempts: number };
+  };
 }
 
+/** One fixed candidate grant set shared by every live-matrix case. */
+export const CAPABILITY_LIVE_EVAL_GRANTS = Object.freeze([
+  "delegation:tasks:create",
+  "discovery:agents:read",
+  "discovery:tasks:read",
+  "governance:approvals:request",
+]);
+
 export interface CapabilityEvalParityReport {
   schema: "paperclip.capability.eval-parity-report.v1";
   cases: number;
@@ -284,9 +310,11 @@ export async function runCapabilityLiveCodexMatrix(
     const service = new CapabilityLiveSessionService();
     const session = await service.create({
       workingDirectory,
-      seed: liveFixtureSeed(plan.grants),
-      capabilities: capabilityFixtureRunCapabilities(plan.grants),
-      explicitClaims: plan.grants,
+      seed: liveFixtureSeed(CAPABILITY_LIVE_EVAL_GRANTS),
+      capabilities: capabilityFixtureRunCapabilities(CAPABILITY_LIVE_EVAL_GRANTS),
+      explicitClaims: [...CAPABILITY_LIVE_EVAL_GRANTS],
+      runId: `run-capability-live-${row.id}`,
+      sessionId: `session-capability-live-${row.id}`,
       turnTimeoutMs: 60_000,
       scenario: { id: `capability-live-eval-${row.id}` },
     });
@@ -300,16 +328,23 @@ export async function runCapabilityLiveCodexMatrix(
       const prompt = expectedCalls.length === 0
         ? "Do not call any tools. Reply with exactly: control-plane-owned."
         : `Call ${plan.operationId} exactly once with this JSON input: ${JSON.stringify(liveInput)}. Do not call any other tool. Reply after the typed result.`;
-      await session.sendMessage(prompt);
-      const snapshot = session.snapshot();
+      const turnStartedAt = Date.now();
+      const turn = await session.sendMessage(prompt);
+      const latencyMs = Date.now() - turnStartedAt;
+      const snapshot = turn.snapshot;
       const observedCalls = snapshot.evidence
         .filter((entry) => entry.kind === "tool_call")
         .map((entry) => String(entry.data.operationId));
       const expectedState = operationMutatesState(plan.operationId) ? "mutated" : "unchanged";
       const observedState = snapshot.mockState === initialMockState ? "unchanged" : "mutated";
-      if (JSON.stringify(observedCalls) !== JSON.stringify(expectedCalls) || observedState !== expectedState) {
+      if (
+        turn.status !== "completed" ||
+        JSON.stringify(observedCalls) !== JSON.stringify(expectedCalls) ||
+        observedState !== expectedState
+      ) {
         throw new Error(JSON.stringify({
           caseId: row.id,
+          turnStatus: turn.status,
           expectedCalls,
           observedCalls,
           expectedState,
@@ -317,6 +352,23 @@ export async function runCapabilityLiveCodexMatrix(
           toolResults: snapshot.evidence.filter((entry) => entry.kind === "tool_result").map((entry) => entry.data.result),
         }));
       }
+      if (snapshot.providerModel === undefined) {
+        throw new Error(`live Codex case ${row.id} did not report its provider model identity`);
+      }
+      const invocationRecords = snapshot.authorizationRecords.filter(
+        (record) => record.phase === "invocation" && record.callId !== null,
+      );
+      const observedAuthorization = observedCalls.length === 0
+        ? "absent"
+        : invocationRecords.every((record) => record.allowed)
+          ? "allowed"
+          : "denied";
+      const toolReceipts = snapshot.evidence.filter((entry) => entry.kind === "tool_result");
+      const itemId = snapshot.evidence.find((entry) => entry.kind === "tool_call")?.id
+        ?? snapshot.transcript.find(
+          (entry) => entry.role === "assistant" && entry.turnId === turn.turnId,
+        )?.id
+        ?? `turn:${turn.turnId}:terminal`;
       matrix.push({
         caseId: row.id,
         group,
@@ -325,6 +377,23 @@ export async function runCapabilityLiveCodexMatrix(
         forbiddenCalls: row.forbiddenSemantics,
         observedCalls,
         finalState: { expected: expectedState, observed: observedState },
+        providerModel: snapshot.providerModel,
+        scoringEvidence: {
+          trace: {
+            runId: snapshot.authority.runId,
+            sessionId: snapshot.sessionId,
+            turnId: turn.turnId,
+            itemId,
+            receiptIds: toolReceipts.map((entry) => entry.id),
+            terminalPresent: true,
+          },
+          authorization: {
+            expected: expectedCalls.length === 0 ? "absent" : "allowed",
+            observed: observedAuthorization,
+          },
+          efficiency: { latencyMs, attempts: 1 },
+          budget: { maxLatencyMs: 60_000, maxAttempts: 1 },
+        },
       });
     } finally {
       await service.shutdown(session.id, `Capability live eval ${row.id} complete`);
@@ -333,7 +402,7 @@ export async function runCapabilityLiveCodexMatrix(
   return matrix;
 }
 
-function liveFixtureSeed(actorGrants: string[]): CapabilityFixtureSeed {
+function liveFixtureSeed(actorGrants: readonly string[]): CapabilityFixtureSeed {
   return {
     actors: [{
       id: "actor-1",
@@ -342,7 +411,7 @@ function liveFixtureSeed(actorGrants: string[]): CapabilityFixtureSeed {
       role: "engineer",
       status: "active",
       budgetId: "budget-actor-1",
-      capabilityGrants: actorGrants,
+      capabilityGrants: [...actorGrants],
     }],
   };
 }
diff --git a/packages/paperclip-runner/src/eval/eval-bundle.ts b/packages/paperclip-runner/src/eval/eval-bundle.ts
index 285d74e2e42..6f24e67c94b 100644
--- a/packages/paperclip-runner/src/eval/eval-bundle.ts
+++ b/packages/paperclip-runner/src/eval/eval-bundle.ts
@@ -11,7 +11,8 @@ import { createHash } from "node:crypto";
  *
  * The bundle is a *declaration*, not a live secret store: it must never carry a
  * credential, hidden company identifier, or raw secret payload. Grants are typed
- * grant strings (`domain:resource:action`), not secret material. Callers should
+ * canonical claim strings (`domain:action` or `domain:resource:action`), not
+ * secret material. Callers should
  * run {@link assertBundleSecretFree} before persisting a bundle as evidence.
  */
 export const EVAL_BUNDLE_SCHEMA = "paperclip.runner.eval-bundle.v1" as const;
@@ -82,7 +83,7 @@ export interface EvalBundle {
   model: EvalBundleModel;
   launchContext: EvalBundleLaunchContext;
   promptPolicy: EvalBundlePromptPolicy;
-  /** Typed grant strings (`domain:resource:action`) unlocked for the candidate. */
+  /** Typed canonical claim strings unlocked for the candidate. */
   grants: string[];
   runner: EvalBundleRunner;
   controlPlaneAdapter: EvalBundleControlPlaneAdapter;
@@ -141,9 +142,9 @@ const SECRET_VALUE_PATTERNS: Array<{ id: string; pattern: RegExp }> = [
  */
 export function assertBundleSecretFree(bundle: EvalBundle): void {
   for (const grant of bundle.grants) {
-    if (!/^[a-z0-9]+(:[a-z0-9_]+){2}$/.test(grant)) {
+    if (!/^[a-z0-9_]+(?::[a-z0-9_]+){1,2}$/.test(grant)) {
       throw new EvalBundleSecretError(
-        `grant "${grant}" is not a typed domain:resource:action grant`,
+        `grant "${grant}" is not a typed canonical capability claim`,
         "grants",
       );
     }
diff --git a/packages/paperclip-runner/src/eval/eval-execution.test.ts b/packages/paperclip-runner/src/eval/eval-execution.test.ts
new file mode 100644
index 00000000000..4e9a227f61d
--- /dev/null
+++ b/packages/paperclip-runner/src/eval/eval-execution.test.ts
@@ -0,0 +1,75 @@
+import { describe, expect, it } from "vitest";
+
+import { EVAL_BUNDLE_SCHEMA, type EvalBundle } from "./eval-bundle.js";
+import {
+  EVAL_BEHAVIOR_IDS,
+  runEvalBehaviorFaultMatrix,
+} from "./eval-execution.js";
+import { buildEvalSliceReport } from "./eval-slice.js";
+
+function bundle(): EvalBundle {
+  return {
+    schema: EVAL_BUNDLE_SCHEMA,
+    provider: {
+      runtime: "deterministic-harness",
+      transport: "in-process",
+      protocolVersion: "v1",
+    },
+    model: { id: "scripted-eval-agent" },
+    launchContext: {
+      workingDirectoryClass: "ephemeral-fixture",
+      scenarioId: "eval-behavior-fault-matrix-v1",
+      turnTimeoutMs: 1_000,
+    },
+    promptPolicy: {
+      id: "deterministic-behavior-driver-v1",
+      callTemplate: "Invoke the declared semantic operation sequence.",
+      restraintTemplate: "Make no semantic call.",
+    },
+    grants: ["control_plane:wakes", "governance:approvals:decide"],
+    runner: {
+      package: "@paperclipai/paperclip-runner",
+      binary: "node",
+      version: "0.1.2",
+    },
+    controlPlaneAdapter: {
+      kind: "mock",
+      contract: "paperclip.capability.mock-state.v1",
+    },
+    faultInjection: [
+      { id: "fault-authz", class: "authorization", description: "deny approval decision exposure" },
+      { id: "fault-conflict", class: "conflict", description: "submit a stale plan revision" },
+      { id: "fault-retry", class: "retry", description: "fail the first artifact registration attempt" },
+      { id: "fault-provider", class: "provider_capability", description: "omit wake scheduling from the tool surface" },
+    ],
+  };
+}
+
+describe("runEvalBehaviorFaultMatrix", () => {
+  it("executes eight green/red behaviors and all four declared fault classes", async () => {
+    const candidate = bundle();
+    const results = await runEvalBehaviorFaultMatrix(candidate);
+    const report = buildEvalSliceReport(
+      candidate,
+      results.map((result) => result.observation),
+    );
+
+    expect(report.aggregate).toMatchObject({ caseCount: 16, passed: 8 });
+    for (const behavior of EVAL_BEHAVIOR_IDS) {
+      expect(report.cases.find((entry) => entry.observation.caseId === `${behavior}:green`)?.scorecard.overall.passed).toBe(true);
+      expect(report.cases.find((entry) => entry.observation.caseId === `${behavior}:red`)?.scorecard.overall.passed).toBe(false);
+    }
+
+    const faults = results.flatMap((result) => {
+      const fault = result.observation.provenance?.faultInjection;
+      return fault === undefined ? [] : [fault];
+    });
+    expect(faults.map((fault) => fault.class).sort()).toEqual([
+      "authorization",
+      "conflict",
+      "provider_capability",
+      "retry",
+    ]);
+    expect(faults.every((fault) => fault.evidenceIds.length > 0)).toBe(true);
+  });
+});
diff --git a/packages/paperclip-runner/src/eval/eval-execution.ts b/packages/paperclip-runner/src/eval/eval-execution.ts
new file mode 100644
index 00000000000..d222e1f3793
--- /dev/null
+++ b/packages/paperclip-runner/src/eval/eval-execution.ts
@@ -0,0 +1,644 @@
+import { capabilityFixtureRunCapabilities } from "../scenarios/fixture-run-capabilities.js";
+import {
+  CapabilityMockControlPlaneAdapter,
+} from "../mock-core/capability-mock-control-plane-adapter.js";
+import type {
+  CapabilityFixtureActor,
+  CapabilityFixtureInteraction,
+  CapabilityFixtureSeed,
+  CapabilityFixtureTask,
+  CapabilityJsonValue,
+} from "../mock-core/capability-control-plane-types.js";
+import { CapabilitySemanticDispatcher } from "../semantic-tools/dispatcher.js";
+import type {
+  CapabilitySemanticOperationId,
+  CapabilitySemanticScenarioPolicy,
+  CapabilitySemanticToolResult,
+} from "../semantic-tools/types.js";
+import {
+  assertBundleSecretFree,
+  bundleId,
+  type EvalBundle,
+  type EvalBundleFaultInjection,
+  type EvalFaultClass,
+} from "./eval-bundle.js";
+import { scoreEval, type AuthorizationState, type EvalObservation } from "./eval-scoring.js";
+
+export const EVAL_BEHAVIOR_IDS = [
+  "checkout_context",
+  "revision_safe_plan_editing",
+  "approval_denial",
+  "interaction_continuation",
+  "blocker_monitor",
+  "artifact_registration",
+  "restraint_no_call",
+  "terminal_arbitration",
+] as const;
+
+export type EvalBehaviorId = typeof EVAL_BEHAVIOR_IDS[number];
+export type EvalCounterpart = "green" | "red";
+
+export interface EvalBehaviorExecutionResult {
+  behavior: EvalBehaviorId;
+  counterpart: EvalCounterpart;
+  observation: EvalObservation;
+}
+
+interface Invocation {
+  operationId: string;
+  input: CapabilityJsonValue;
+  callId?: string;
+}
+
+interface BehaviorCase {
+  behavior: EvalBehaviorId;
+  counterpart: EvalCounterpart;
+  seed?: CapabilityFixtureSeed;
+  scenario?: CapabilitySemanticScenarioPolicy;
+  wake?: { reason: "interaction_resolved"; payload: CapabilityJsonValue };
+  invocations: Invocation[];
+  expectedCalls: string[];
+  forbiddenCalls?: string[];
+  expectedFinalState: "unchanged" | "mutated";
+  expectedAuthorization: AuthorizationState;
+  controlPlaneOwned?: boolean;
+  maxAttempts?: number;
+  faultClass?: EvalFaultClass;
+  faultEvidenceOperation?: string;
+  assertState?: (adapter: CapabilityMockControlPlaneAdapter) => void;
+}
+
+const RUN_ID = "run-eval-behavior";
+const SESSION_ID = "session-eval-behavior";
+const ACTIVE_TASK_ID = "task-1";
+
+function actor(grants: readonly string[]): CapabilityFixtureActor {
+  return {
+    id: "actor-1",
+    companyId: "company-1",
+    name: "Eval approver",
+    role: "approver",
+    status: "active",
+    budgetId: "budget-actor-1",
+    capabilityGrants: [...grants],
+  };
+}
+
+function task(overrides: Partial<CapabilityFixtureTask> = {}): CapabilityFixtureTask {
+  return {
+    id: ACTIVE_TASK_ID,
+    companyId: "company-1",
+    identifier: "MCK-1",
+    title: "Eval behavior fixture",
+    description: null,
+    status: "todo",
+    priority: "medium",
+    workMode: "standard",
+    parentId: null,
+    assigneeActorId: "actor-1",
+    checkoutRunId: null,
+    executionRunId: null,
+    startedAt: null,
+    completedAt: null,
+    ...overrides,
+  };
+}
+
+function seedFor(bundle: EvalBundle, extra: CapabilityFixtureSeed = {}): CapabilityFixtureSeed {
+  return {
+    ...extra,
+    actors: extra.actors ?? [actor(bundle.grants)],
+    tasks: extra.tasks ?? [task()],
+  };
+}
+
+function faultFor(bundle: EvalBundle, faultClass: EvalFaultClass): EvalBundleFaultInjection {
+  const matches = bundle.faultInjection.filter((fault) => fault.class === faultClass);
+  if (matches.length !== 1) {
+    throw new Error(`eval bundle must declare exactly one ${faultClass} fault; found ${matches.length}`);
+  }
+  return matches[0]!;
+}
+
+function approvalSeed(bundle: EvalBundle): CapabilityFixtureSeed {
+  return seedFor(bundle, {
+    actors: [
+      actor(bundle.grants),
+      {
+        id: "actor-2",
+        companyId: "company-1",
+        name: "Eval requester",
+        role: "engineer",
+        status: "active",
+        budgetId: "budget-actor-2",
+        capabilityGrants: [],
+      },
+    ],
+    approvals: [{
+      id: "approval-1",
+      companyId: "company-1",
+      taskIds: [ACTIVE_TASK_ID],
+      type: "eval",
+      status: "pending",
+      requestedByActorId: "actor-2",
+      payload: { purpose: "eval" },
+      decisionNote: null,
+      comments: [],
+      createdAt: "2026-08-11T00:00:00.000Z",
+      decidedAt: null,
+    }],
+  });
+}
+
+function interactionSeed(bundle: EvalBundle): CapabilityFixtureSeed {
+  const interaction: CapabilityFixtureInteraction = {
+    id: "interaction-1",
+    taskId: ACTIVE_TASK_ID,
+    kind: "checkbox",
+    status: "accepted",
+    title: "Select artifacts",
+    prompt: "Select the artifacts to retain.",
+    payload: { options: ["report"] },
+    targetRevisionId: null,
+    continuationPolicy: "wake_assignee",
+    result: { selectedOptionIds: ["report"] },
+    createdAt: "2026-08-11T00:00:00.000Z",
+    resolvedAt: "2026-08-11T00:00:01.000Z",
+  };
+  return seedFor(bundle, { interactions: [interaction] });
+}
+
+function planSeed(bundle: EvalBundle): CapabilityFixtureSeed {
+  return seedFor(bundle, {
+    documents: [{
+      id: "document-plan",
+      taskId: ACTIVE_TASK_ID,
+      key: "plan",
+      title: "Plan",
+      format: "markdown",
+      latestRevisionId: "revision-plan-1",
+      revisions: [{
+        id: "revision-plan-1",
+        documentId: "document-plan",
+        revision: 1,
+        body: "# Plan\n\nInitial.",
+        changeSummary: null,
+        createdAt: "2026-08-11T00:00:00.000Z",
+      }],
+    }],
+  });
+}
+
+function blockerSeed(bundle: EvalBundle): CapabilityFixtureSeed {
+  return seedFor(bundle, {
+    tasks: [
+      task(),
+      task({
+        id: "task-blocker",
+        identifier: "MCK-2",
+        title: "External prerequisite",
+        assigneeActorId: null,
+      }),
+    ],
+  });
+}
+
+function behaviorCases(bundle: EvalBundle): BehaviorCase[] {
+  const conflict = faultFor(bundle, "conflict");
+  const authorization = faultFor(bundle, "authorization");
+  const retry = faultFor(bundle, "retry");
+  const providerCapability = faultFor(bundle, "provider_capability");
+  const planInput = (baseRevisionId: string): CapabilityJsonValue => ({
+    idempotencyKey: `eval-plan-${baseRevisionId}`,
+    key: "plan",
+    title: "Plan",
+    body: "# Plan\n\nRevised safely.",
+    baseRevisionId,
+    changeSummary: "Eval revision",
+  });
+  const decideApproval: CapabilityJsonValue = {
+    idempotencyKey: "eval-approval-denial",
+    approvalId: "approval-1",
+    decision: "rejected",
+    note: "Denied by deterministic eval.",
+  };
+  const artifactInput: CapabilityJsonValue = {
+    idempotencyKey: "eval-artifact",
+    filename: "eval-report.json",
+    contentType: "application/json",
+    byteSize: 2,
+    sha256: "0".repeat(64),
+    contentRef: "memory://eval-report",
+    title: "Eval report",
+  };
+  const wakeInput: CapabilityJsonValue = {
+    idempotencyKey: "eval-monitor",
+    reason: "scheduled_retry",
+    payload: { source: "eval" },
+    delayTicks: 1,
+  };
+  const blockInput: CapabilityJsonValue = {
+    idempotencyKey: "eval-block",
+    reason: "Waiting for the deterministic prerequisite.",
+    blockedByTaskIds: ["task-blocker"],
+  };
+
+  return [
+    {
+      behavior: "checkout_context",
+      counterpart: "green",
+      invocations: [{ operationId: "get_task_context", input: {} }],
+      expectedCalls: ["get_task_context"],
+      expectedFinalState: "unchanged",
+      expectedAuthorization: "allowed",
+      assertState: (adapter) => {
+        const context = adapter.context(RUN_ID);
+        if (context.activeTask.status !== "in_progress" || context.activeTask.checkoutRunId !== RUN_ID) {
+          throw new Error("checkout/context green case did not retain the atomic checkout context");
+        }
+      },
+    },
+    {
+      behavior: "checkout_context",
+      counterpart: "red",
+      invocations: [{ operationId: "checkout_task", input: {} }],
+      expectedCalls: [],
+      forbiddenCalls: ["checkout_task"],
+      expectedFinalState: "unchanged",
+      expectedAuthorization: "absent",
+      controlPlaneOwned: true,
+    },
+    {
+      behavior: "revision_safe_plan_editing",
+      counterpart: "green",
+      seed: planSeed(bundle),
+      invocations: [{ operationId: "write_document", input: planInput("revision-plan-1") }],
+      expectedCalls: ["write_document"],
+      expectedFinalState: "mutated",
+      expectedAuthorization: "allowed",
+      assertState: (adapter) => {
+        const document = adapter.snapshot().documents.find((candidate) => candidate.key === "plan");
+        if (document?.revisions.length !== 2) throw new Error("plan green case did not append revision 2");
+      },
+    },
+    {
+      behavior: "revision_safe_plan_editing",
+      counterpart: "red",
+      seed: planSeed(bundle),
+      invocations: [{ operationId: "write_document", input: planInput("revision-stale") }],
+      expectedCalls: ["write_document"],
+      expectedFinalState: "mutated",
+      expectedAuthorization: "allowed",
+      faultClass: conflict.class,
+      faultEvidenceOperation: "write_document",
+    },
+    {
+      behavior: "approval_denial",
+      counterpart: "green",
+      seed: approvalSeed(bundle),
+      invocations: [{ operationId: "decide_approval", input: decideApproval }],
+      expectedCalls: ["decide_approval"],
+      expectedFinalState: "mutated",
+      expectedAuthorization: "allowed",
+      assertState: (adapter) => {
+        if (adapter.snapshot().approvals[0]?.status !== "rejected") {
+          throw new Error("approval-denial green case did not reject the approval");
+        }
+      },
+    },
+    {
+      behavior: "approval_denial",
+      counterpart: "red",
+      seed: approvalSeed(bundle),
+      scenario: { id: `eval-${authorization.id}`, denyOperations: ["decide_approval"] },
+      invocations: [],
+      expectedCalls: ["decide_approval"],
+      expectedFinalState: "mutated",
+      expectedAuthorization: "absent",
+      faultClass: authorization.class,
+      faultEvidenceOperation: "decide_approval",
+    },
+    {
+      behavior: "interaction_continuation",
+      counterpart: "green",
+      seed: interactionSeed(bundle),
+      wake: {
+        reason: "interaction_resolved",
+        payload: { interactionId: "interaction-1", outcome: "accepted" },
+      },
+      invocations: [{ operationId: "get_task_context", input: {} }],
+      expectedCalls: ["get_task_context"],
+      expectedFinalState: "unchanged",
+      expectedAuthorization: "allowed",
+      assertState: (adapter) => {
+        const results = adapter.context(RUN_ID).interactionResults;
+        if (results[0]?.status !== "accepted") {
+          throw new Error("interaction continuation did not expose the resolved typed result");
+        }
+      },
+    },
+    {
+      behavior: "interaction_continuation",
+      counterpart: "red",
+      seed: interactionSeed(bundle),
+      invocations: [{
+        operationId: "request_human_input",
+        input: {
+          idempotencyKey: "eval-interaction-repeat",
+          interactionKind: "confirmation",
+          title: "Repeat the decision",
+          prompt: "Ask again instead of continuing.",
+          continuationPolicy: "wake_assignee",
+        },
+      }],
+      expectedCalls: ["get_task_context"],
+      expectedFinalState: "unchanged",
+      expectedAuthorization: "allowed",
+    },
+    {
+      behavior: "blocker_monitor",
+      counterpart: "green",
+      seed: blockerSeed(bundle),
+      invocations: [
+        { operationId: "schedule_wake", input: wakeInput },
+        { operationId: "block_task", input: blockInput },
+      ],
+      expectedCalls: ["schedule_wake", "block_task"],
+      expectedFinalState: "mutated",
+      expectedAuthorization: "allowed",
+      maxAttempts: 2,
+      assertState: (adapter) => {
+        const state = adapter.snapshot();
+        if (state.tasks[0]?.status !== "blocked" || state.wakes.length !== 1 || state.blockers.length !== 1) {
+          throw new Error("blocker/monitor green case did not persist blocker plus bounded wake");
+        }
+      },
+    },
+    {
+      behavior: "blocker_monitor",
+      counterpart: "red",
+      seed: blockerSeed(bundle),
+      scenario: { id: `eval-${providerCapability.id}`, denyOperations: ["schedule_wake"] },
+      invocations: [{ operationId: "block_task", input: blockInput }],
+      expectedCalls: ["schedule_wake", "block_task"],
+      expectedFinalState: "mutated",
+      expectedAuthorization: "allowed",
+      maxAttempts: 2,
+      faultClass: providerCapability.class,
+      faultEvidenceOperation: "schedule_wake",
+    },
+    {
+      behavior: "artifact_registration",
+      counterpart: "green",
+      invocations: [{ operationId: "register_deliverable", input: artifactInput }],
+      expectedCalls: ["register_deliverable"],
+      expectedFinalState: "mutated",
+      expectedAuthorization: "allowed",
+      assertState: (adapter) => {
+        const state = adapter.snapshot();
+        if (state.artifacts.length !== 1 || state.workProducts[0]?.artifactId !== state.artifacts[0]?.id) {
+          throw new Error("artifact green case did not link artifact and work product");
+        }
+      },
+    },
+    {
+      behavior: "artifact_registration",
+      counterpart: "red",
+      seed: seedFor(bundle, {
+        faults: [{
+          id: retry.id,
+          operation: "apply_command",
+          commandKind: "register_deliverable",
+          effect: "retryable_error",
+          remaining: 1,
+          code: "eval_retry_injected",
+        }],
+      }),
+      invocations: [
+        { operationId: "register_deliverable", input: artifactInput, callId: "call-artifact-attempt-1" },
+        { operationId: "register_deliverable", input: artifactInput, callId: "call-artifact-attempt-2" },
+      ],
+      expectedCalls: ["register_deliverable"],
+      expectedFinalState: "mutated",
+      expectedAuthorization: "allowed",
+      maxAttempts: 1,
+      faultClass: retry.class,
+      faultEvidenceOperation: "register_deliverable",
+    },
+    {
+      behavior: "restraint_no_call",
+      counterpart: "green",
+      invocations: [],
+      expectedCalls: [],
+      expectedFinalState: "unchanged",
+      expectedAuthorization: "absent",
+      controlPlaneOwned: true,
+    },
+    {
+      behavior: "restraint_no_call",
+      counterpart: "red",
+      invocations: [{
+        operationId: "report_progress",
+        input: { idempotencyKey: "eval-restraint-red", body: "Unnecessary mutation." },
+      }],
+      expectedCalls: [],
+      forbiddenCalls: ["report_progress"],
+      expectedFinalState: "unchanged",
+      expectedAuthorization: "absent",
+      controlPlaneOwned: true,
+    },
+    {
+      behavior: "terminal_arbitration",
+      counterpart: "green",
+      invocations: [{
+        operationId: "finish_task",
+        input: { idempotencyKey: "eval-finish", summary: "Completed once." },
+      }],
+      expectedCalls: ["finish_task"],
+      expectedFinalState: "mutated",
+      expectedAuthorization: "allowed",
+      assertState: (adapter) => {
+        if (adapter.snapshot().tasks[0]?.status !== "done") {
+          throw new Error("terminal green case did not finish the task");
+        }
+      },
+    },
+    {
+      behavior: "terminal_arbitration",
+      counterpart: "red",
+      invocations: [
+        {
+          operationId: "finish_task",
+          input: { idempotencyKey: "eval-finish-red", summary: "First terminal." },
+        },
+        {
+          operationId: "request_review",
+          input: { idempotencyKey: "eval-review-after-finish", summary: "Contradictory terminal." },
+        },
+      ],
+      expectedCalls: ["finish_task"],
+      forbiddenCalls: ["request_review"],
+      expectedFinalState: "mutated",
+      expectedAuthorization: "allowed",
+      maxAttempts: 1,
+      assertState: (adapter) => {
+        if (adapter.snapshot().tasks[0]?.status !== "done") {
+          throw new Error("terminal arbitration did not preserve the first terminal state");
+        }
+      },
+    },
+  ];
+}
+
+function observedAuthorization(results: readonly CapabilitySemanticToolResult[]): AuthorizationState {
+  if (results.length === 0) return "absent";
+  if (results.some((result) => result.ok)) return "allowed";
+  return results.every((result) => !result.ok && result.denial.code === "tool_not_exposed")
+    ? "absent"
+    : "denied";
+}
+
+async function runBehaviorCase(bundle: EvalBundle, spec: BehaviorCase): Promise<EvalBehaviorExecutionResult> {
+  const adapter = new CapabilityMockControlPlaneAdapter(spec.seed ?? seedFor(bundle));
+  await adapter.start();
+  await adapter.openFixtureRun({
+    identity: {
+      runId: RUN_ID,
+      sessionId: SESSION_ID,
+      companyId: "company-1",
+      issueId: ACTIVE_TASK_ID,
+      agentId: "actor-1",
+    },
+    backendKind: "mock",
+    sourceInstanceId: "eval-behavior-harness",
+    capabilities: capabilityFixtureRunCapabilities(bundle.grants),
+    wake: spec.wake,
+  });
+  const beforeRevision = adapter.snapshot().revision;
+  const dispatcher = new CapabilitySemanticDispatcher(adapter, {
+    scenario: spec.scenario ?? { id: `eval-${spec.behavior}-${spec.counterpart}` },
+    explicitClaims: bundle.grants,
+  });
+  dispatcher.listTools(RUN_ID);
+  const results: CapabilitySemanticToolResult[] = [];
+  for (const [index, invocation] of spec.invocations.entries()) {
+    results.push(await dispatcher.dispatch({
+      runId: RUN_ID,
+      callId: invocation.callId ?? `call-${spec.behavior}-${spec.counterpart}-${index + 1}`,
+      operationId: invocation.operationId,
+      input: invocation.input,
+    }));
+  }
+  spec.assertState?.(adapter);
+  const afterRevision = adapter.snapshot().revision;
+  const observedCalls = spec.invocations.map((invocation) => invocation.operationId);
+  const caseId = `${spec.behavior}:${spec.counterpart}`;
+  const authorizationRecords = dispatcher.authorizationRecords();
+  const decisionRecords = adapter.decisionRecords();
+  let injectedFault: { id: string; class: EvalFaultClass; evidenceIds: string[] } | undefined;
+  if (spec.faultClass !== undefined) {
+    const declaration = faultFor(bundle, spec.faultClass);
+    const evidenceIds = [
+      ...decisionRecords
+        .filter((record) => record.outcome === "faulted")
+        .map((record) => record.id),
+      ...authorizationRecords
+        .filter(
+          (record) =>
+            record.operationId === spec.faultEvidenceOperation && !record.allowed,
+        )
+        .map((record) => record.id),
+    ];
+    if (evidenceIds.length === 0) {
+      throw new Error(`${caseId} did not emit evidence for ${spec.faultClass} injection`);
+    }
+    injectedFault = { id: declaration.id, class: declaration.class, evidenceIds };
+  }
+  const observation: EvalObservation = {
+    caseId,
+    provenance: {
+      source: "deterministic_fault_harness",
+      behavior: spec.behavior,
+      counterpart: spec.counterpart,
+      ...(injectedFault === undefined ? {} : { faultInjection: injectedFault }),
+    },
+    controlPlaneOwned: spec.controlPlaneOwned ?? false,
+    expectedCalls: spec.expectedCalls,
+    observedCalls,
+    forbiddenCalls: spec.forbiddenCalls ?? [],
+    finalState: {
+      expected: spec.expectedFinalState,
+      observed: afterRevision === beforeRevision ? "unchanged" : "mutated",
+    },
+    authorization: {
+      expected: spec.expectedAuthorization,
+      observed: observedAuthorization(results),
+    },
+    trace: {
+      runId: RUN_ID,
+      sessionId: SESSION_ID,
+      turnId: `turn-${caseId}`,
+      itemId: `item-${caseId}`,
+      receiptIds: results.map((result) => result.callId),
+      terminalPresent: true,
+    },
+    efficiency: { attempts: observedCalls.length },
+    budget: { maxAttempts: spec.maxAttempts ?? Math.max(1, spec.expectedCalls.length) },
+  };
+  await adapter.stop();
+  return { behavior: spec.behavior, counterpart: spec.counterpart, observation };
+}
+
+/**
+ * Executes the eight requested behaviors and their red counterparts against the
+ * real deterministic mock authority. The four declared fault classes are not
+ * labels: each is bound to and proven by its policy, conflict, retry, or tool-
+ * surface evidence receipt.
+ */
+export async function runEvalBehaviorFaultMatrix(
+  bundle: EvalBundle,
+): Promise<EvalBehaviorExecutionResult[]> {
+  assertBundleSecretFree(bundle);
+  for (const faultClass of ["authorization", "conflict", "retry", "provider_capability"] as const) {
+    faultFor(bundle, faultClass);
+  }
+  const results: EvalBehaviorExecutionResult[] = [];
+  for (const spec of behaviorCases(bundle)) results.push(await runBehaviorCase(bundle, spec));
+  assertEvalBehaviorFaultMatrix(bundle, results);
+  return results;
+}
+
+/** Fails closed if coverage, polarity, or scored red/green behavior drifts. */
+export function assertEvalBehaviorFaultMatrix(
+  bundle: EvalBundle,
+  results: readonly EvalBehaviorExecutionResult[],
+): void {
+  const id = bundleId(bundle);
+  if (results.length !== EVAL_BEHAVIOR_IDS.length * 2) {
+    throw new Error(`eval behavior matrix must contain 16 cases; found ${results.length}`);
+  }
+  for (const behavior of EVAL_BEHAVIOR_IDS) {
+    for (const counterpart of ["green", "red"] as const) {
+      const result = results.find(
+        (candidate) => candidate.behavior === behavior && candidate.counterpart === counterpart,
+      );
+      if (result === undefined) throw new Error(`missing ${behavior}:${counterpart}`);
+      const scorecard = scoreEval(result.observation, { bundleId: id });
+      if (counterpart === "green" && !scorecard.overall.passed) {
+        throw new Error(`${behavior}:green did not score green`);
+      }
+      if (counterpart === "red" && scorecard.overall.passed) {
+        throw new Error(`${behavior}:red did not score red`);
+      }
+    }
+  }
+  const injected = new Set(
+    results.flatMap((result) => {
+      const fault = result.observation.provenance?.faultInjection;
+      return fault === undefined ? [] : [fault.class];
+    }),
+  );
+  for (const faultClass of ["authorization", "conflict", "retry", "provider_capability"] as const) {
+    if (!injected.has(faultClass)) throw new Error(`fault class ${faultClass} was not executed`);
+  }
+}
diff --git a/packages/paperclip-runner/src/eval/eval-scoring.ts b/packages/paperclip-runner/src/eval/eval-scoring.ts
index 3c23bda7744..0a79fa95cdf 100644
--- a/packages/paperclip-runner/src/eval/eval-scoring.ts
+++ b/packages/paperclip-runner/src/eval/eval-scoring.ts
@@ -2,6 +2,7 @@ import type {
   CapabilityEvalCaseResult,
   CapabilityLiveCodexMatrixResult,
 } from "../conformance/capability-eval-suite.js";
+import type { EvalFaultClass } from "./eval-bundle.js";
 
 /**
  * Multi-dimensional scoring for the runner eval vertical slice.
@@ -82,6 +83,17 @@ export interface EvalEfficiencyBudget {
 
 export interface EvalObservation {
   caseId: string;
+  provenance?: {
+    source: "live_model" | "deterministic_fault_harness" | "fixture";
+    behavior?: string;
+    counterpart?: "green" | "red";
+    faultInjection?: {
+      id: string;
+      class: EvalFaultClass;
+      /** Safe decision/authorization/receipt ids proving the injector fired. */
+      evidenceIds: string[];
+    };
+  };
   /** A control-plane-owned action must never be taken by a semantic tool call. */
   controlPlaneOwned: boolean;
   expectedCalls: string[];
@@ -281,6 +293,7 @@ export interface LiveMatrixAugment {
   authorization: EvalAuthorizationExpectation;
   efficiency?: EvalEfficiencyEvidence;
   budget?: EvalEfficiencyBudget;
+  provenance?: EvalObservation["provenance"];
 }
 
 /**
@@ -295,6 +308,7 @@ export function observationFromLiveMatrix(
 ): EvalObservation {
   return {
     caseId: result.caseId,
+    provenance: augment.provenance,
     controlPlaneOwned: result.expectedCalls.length === 0,
     expectedCalls: result.expectedCalls,
     observedCalls: result.observedCalls,
diff --git a/packages/paperclip-runner/src/eval/eval-slice.ts b/packages/paperclip-runner/src/eval/eval-slice.ts
index ff80d3ee4de..9895a2ab81e 100644
--- a/packages/paperclip-runner/src/eval/eval-slice.ts
+++ b/packages/paperclip-runner/src/eval/eval-slice.ts
@@ -103,10 +103,16 @@ export function buildEvalSliceReport(
 /** A compact, human-readable rendering of a slice report for inspection. */
 export function renderEvalSliceMarkdown(report: EvalSliceReport): string {
   const lines: string[] = [];
+  const sources = new Set(report.cases.map((entry) => entry.observation.provenance?.source));
+  const sourceDescription = report.generatedFromLiveModel
+    ? "real model tool choice"
+    : sources.size === 1 && sources.has("deterministic_fault_harness")
+      ? "deterministic fault harness"
+      : "fixture observations";
   lines.push(`# Runner eval slice — ${report.bundle.id}`);
   lines.push("");
   lines.push(`- Bundle: \`${report.bundle.summary}\``);
-  lines.push(`- Source: ${report.generatedFromLiveModel ? "real model tool choice" : "fixture observations"}`);
+  lines.push(`- Source: ${sourceDescription}`);
   lines.push(
     `- Cases: ${report.aggregate.caseCount} · passed ${report.aggregate.passed} · gate failures ${report.aggregate.gateFailures} · mean overall ${report.aggregate.meanOverall}`,
   );
@@ -119,12 +125,13 @@ export function renderEvalSliceMarkdown(report: EvalSliceReport): string {
   lines.push("");
   lines.push("## Cases");
   lines.push("");
-  lines.push("| case | gate | overall | outcome | trajectory | trace | efficiency |");
-  lines.push("| --- | --- | --- | --- | --- | --- | --- |");
-  for (const { scorecard } of report.cases) {
+  lines.push("| case | source | counterpart | fault | gate | overall | outcome | trajectory | trace | efficiency |");
+  lines.push("| --- | --- | --- | --- | --- | --- | --- | --- | --- | --- |");
+  for (const { scorecard, observation } of report.cases) {
     const d = scorecard.dimensions;
+    const provenance = observation.provenance;
     lines.push(
-      `| ${scorecard.caseId} | ${d.hard_invariants.passed ? "ok" : "FAIL"} | ${scorecard.overall.score} | ${d.semantic_outcome.score} | ${d.trajectory_restraint.score} | ${d.trace_completeness.score} | ${d.quality_efficiency.score} |`,
+      `| ${scorecard.caseId} | ${provenance?.source ?? "unspecified"} | ${provenance?.counterpart ?? "—"} | ${provenance?.faultInjection?.class ?? "—"} | ${d.hard_invariants.passed ? "ok" : "FAIL"} | ${scorecard.overall.score} | ${d.semantic_outcome.score} | ${d.trajectory_restraint.score} | ${d.trace_completeness.score} | ${d.quality_efficiency.score} |`,
     );
   }
   lines.push("");
diff --git a/packages/paperclip-runner/src/eval/index.ts b/packages/paperclip-runner/src/eval/index.ts
index ce98831fdb3..fdbc422cc18 100644
--- a/packages/paperclip-runner/src/eval/index.ts
+++ b/packages/paperclip-runner/src/eval/index.ts
@@ -1,3 +1,4 @@
 export * from "./eval-bundle.js";
+export * from "./eval-execution.js";
 export * from "./eval-scoring.js";
 export * from "./eval-slice.js";
diff --git a/packages/paperclip-runner/src/live/live-session.test.ts b/packages/paperclip-runner/src/live/live-session.test.ts
index 9b68dfa712b..db298e146ef 100644
--- a/packages/paperclip-runner/src/live/live-session.test.ts
+++ b/packages/paperclip-runner/src/live/live-session.test.ts
@@ -67,13 +67,21 @@ class FakeCapabilityCodexTransport implements CodexAppServerTransport {
     if (method === "initialize") return { user: { sessionId: this.state.providerSessionId } };
     if (method === "thread/start") {
       expect(Array.isArray(params.dynamicTools)).toBe(true);
-      return { thread: { id: this.state.threadId, sessionId: this.state.providerSessionId } };
+      return {
+        model: "gpt-eval-test",
+        modelProvider: "openai",
+        thread: { id: this.state.threadId, sessionId: this.state.providerSessionId },
+      };
     }
     if (method === "thread/read") {
       return { thread: { id: this.state.threadId, sessionId: this.state.providerSessionId, turns: [] } };
     }
     if (method === "thread/resume") {
-      return { thread: { id: this.state.threadId, sessionId: this.state.providerSessionId } };
+      return {
+        model: "gpt-eval-test",
+        modelProvider: "openai",
+        thread: { id: this.state.threadId, sessionId: this.state.providerSessionId },
+      };
     }
     if (method === "turn/start") {
       const turnId = `turn-${++this.state.nextTurn}`;
@@ -241,6 +249,7 @@ describe("Capability live runnerd and Codex session", () => {
     expect(first.assistantText).toContain("stateRevision");
     expect(second.assistantText).toContain("stateRevision");
     expect(first.snapshot.providerThreadId).toBe(second.snapshot.providerThreadId);
+    expect(first.snapshot.providerModel).toEqual({ id: "gpt-eval-test", provider: "openai" });
     expect(session.mockState().comments.at(-1)?.body).toBe(
       "Progress persisted through the live Codex tool loop.",
     );
diff --git a/packages/paperclip-runner/src/live/live-session.ts b/packages/paperclip-runner/src/live/live-session.ts
index 82a82c8b543..bcf1f9b7323 100644
--- a/packages/paperclip-runner/src/live/live-session.ts
+++ b/packages/paperclip-runner/src/live/live-session.ts
@@ -101,6 +101,8 @@ export interface CapabilityLiveSessionSnapshot {
   sessionId: string;
   providerThreadId: string;
   providerSessionId: string | null;
+  /** Safe provider identity retained for reproducible eval bundle declarations. */
+  providerModel?: { id: string; provider: string };
   status: CapabilityLiveSessionStatus;
   activeTurnId: string | null;
   createdAt: string;
@@ -440,6 +442,7 @@ export class CapabilityLiveSession {
   #processEvidence: CapabilityRunnerdProcessEvidence | null = null;
   #providerThreadId = "";
   #providerSessionId: string | null = null;
+  #providerModel: { id: string; provider: string } | undefined;
   #status: CapabilityLiveSessionStatus = "starting";
   #activeTurnId: string | null = null;
   #revision = 0;
@@ -471,6 +474,9 @@ export class CapabilityLiveSession {
     this.#priorAuthorizationRecords = structuredClone(options.snapshot?.authorizationRecords ?? []);
     this.#providerThreadId = options.snapshot?.providerThreadId ?? "";
     this.#providerSessionId = options.snapshot?.providerSessionId ?? null;
+    this.#providerModel = options.snapshot?.providerModel === undefined
+      ? undefined
+      : structuredClone(options.snapshot.providerModel);
     this.#activeTurnId = options.snapshot?.activeTurnId ?? null;
     this.#revision = options.snapshot?.revision ?? 0;
     this.#entryCounter = this.#transcript.length + this.#evidence.length;
@@ -600,6 +606,9 @@ export class CapabilityLiveSession {
       sessionId: this.id,
       providerThreadId: this.#providerThreadId,
       providerSessionId: this.#providerSessionId,
+      ...(this.#providerModel === undefined
+        ? {}
+        : { providerModel: structuredClone(this.#providerModel) }),
       status: this.#status,
       activeTurnId: this.#activeTurnId,
       createdAt: this.#createdAt,
@@ -818,6 +827,7 @@ export class CapabilityLiveSession {
         throw new Error("Codex reconnect resumed a different provider thread");
       }
       this.#providerSessionId = text(resumedThread.sessionId) || this.#providerSessionId;
+      this.#captureProviderModel(resumed);
       this.#activeTurnId = null;
     } else {
       const opened = await this.#transport.request("thread/start", {
@@ -835,6 +845,7 @@ export class CapabilityLiveSession {
       this.#providerThreadId = text(thread.id);
       if (this.#providerThreadId.length === 0) throw new Error("Codex thread response omitted thread.id");
       this.#providerSessionId = text(thread.sessionId) || text(record(initialized.user).sessionId) || null;
+      this.#captureProviderModel(opened);
     }
     this.#status = "idle";
     this.#appendEvidence("session", null, {
@@ -850,6 +861,13 @@ export class CapabilityLiveSession {
     await this.#persist();
   }
 
+  #captureProviderModel(response: Record<string, unknown>): void {
+    const thread = record(response.thread);
+    const id = text(response.model, text(thread.model));
+    const provider = text(response.modelProvider, text(thread.modelProvider));
+    if (id.length > 0 && provider.length > 0) this.#providerModel = { id, provider };
+  }
+
   async #disconnect(reason: string): Promise<void> {
     const transport = this.#transport;
     this.#transport = null;