test(runner): finish phase 6 conformance

Exercise all atomic-liveness failpoints, make native cancellation retry-safe, and align the allowlist and verification evidence with the final database-backed matrix.

Co-Authored-By: Paperclip <noreply@paperclip.ing>

packages/paperclip-runner/docs/design/phase-6-thin-paperclip-adapter.md

@@ -703,14 +703,18 @@ from the provisional list above. They are added without expanding authority:
 ```text
 packages/paperclip-runner/src/native-session-runtime.ts
 packages/paperclip-runner/src/native-session-runtime.test.ts
+packages/paperclip-runner/spec/fixtures/status-authority-phase5.json
 server/src/services/native-runtime/native-session-executor.ts
 server/src/services/native-runtime/native-session-executor.test.ts
+server/src/services/native-runtime/evidence-classifier.test.ts
 server/src/services/native-runtime/paperclip-control-plane-port.test.ts
 server/src/services/native-runtime/status-arbiter.test.ts
 packages/paperclip-runner/docs/design/phase-6-thin-paperclip-adapter.md
 ```
 
-The package runtime files own persisted provider-session recovery. The server
+The package runtime files own persisted provider-session recovery. The corpus
+edit moves two existing coverage labels between fixtures without changing any
+fixture input or expected outcome, eliminating an unjoinable fixture. The server
 executor files own only the run-scoped cancellation handle and coordinator
 lease around that package runtime. The colocated port and arbiter tests exercise
 the same approved database and pure-policy seams exposed through the named

packages/paperclip-runner/docs/tutorials/phase-06-thin-paperclip-adapter.md

@@ -57,7 +57,7 @@ pnpm --filter @paperclipai/server exec vitest run \
 ```
 
 This runs the unchanged package conformance suite against the real Paperclip
-port. Twenty-five targeted tests check company binding, duplicate/gap replay,
+port. Twenty-eight targeted tests check company binding, duplicate/gap replay,
 checkpoint recovery, immutable result ingestion, durable-evidence authority,
 workspace-gated finalization, atomic liveness rollback, status-version CAS,
 migration repair, legacy byte equivalence, and audit output. All database tests

packages/paperclip-runner/knowledge/evidence/2026-08-09-phase-06-verification.md

@@ -39,7 +39,7 @@ existing accepted result event, and appended only the missing terminal fact.
 
 # Real Paperclip adapter and public-session task
 
-The complete acceptance gate below passed 25 tests in ten files against
+The complete acceptance gate below passed 28 tests in ten files against
 embedded PostgreSQL with zero skipped tests. Database-backed tests start their
 own database instead of skipping when a developer `DATABASE_URL` is absent.
 The same package conformance suite passed unchanged against
@@ -68,10 +68,11 @@ The focused corpus also proved:
 - a failed workspace barrier preserves the accepted semantic result, leaves the
   issue status unchanged, and records a leased `retryable_failure` plus recovery
   owner/action;
-- four injected failures prove interactions, wakes, blocker bindings, status,
+- six injected failures prove governance binding, interactions, wakes, blocker
+  bindings, recovery, status,
   decisions, and effects roll back together before retry ownership is recorded.
 
-The acceptance-matrix entry points also passed as one 10-file, 25-test gate:
+The acceptance-matrix entry points also passed as one 10-file, 28-test gate:
 
 ```sh
 pnpm --filter @paperclipai/server exec vitest run \

packages/paperclip-runner/knowledge/journal/2026-08-09-phase-06-implementation.md

@@ -47,8 +47,8 @@ existing server authority and keep legacy execution as the default.
 - [Phase 6 reference](../../docs/phase-06-thin-paperclip-adapter.md)
 - [Runnable tutorial](../../docs/tutorials/phase-06-thin-paperclip-adapter.md)
 - Package conformance and recovery: four files, six tests passed.
-- Phase 6 acceptance-matrix entry points: ten files, twenty-five tests passed
-  with zero skips, including the database-backed selected-task canary, four
+- Phase 6 acceptance-matrix entry points: ten files, twenty-eight tests passed
+  with zero skips, including the database-backed selected-task canary, six
   atomic-liveness failpoints, migration, sequencing, bounded recovery, and
   legacy compatibility.
 - Focused database rehearsal: five files, eleven tests passed with zero skips;

server/src/__tests__/native-run-finalizer.test.ts

@@ -1,3 +1,4 @@
 // Acceptance-matrix entry point for evidence, finalization, CAS, and atomic
 // status/liveness effects exercised by the real database-backed port suite.
 import "../services/native-runtime/paperclip-control-plane-port.test.js";
+import "../services/native-runtime/evidence-classifier.test.js";

server/src/services/native-runtime/native-session-executor.test.ts

@@ -96,4 +96,23 @@ describe("native session cancellation", () => {
     await running;
     await expect(cancelNativeSession(execution.binding.runId, "late cancel")).resolves.toBe(false);
   });
+
+  it("allows cancellation to be retried when the session dispatch fails", async () => {
+    state.cancel.mockRejectedValueOnce(new Error("transport unavailable"));
+    const running = executePaperclipNativeSession({
+      db: leaseDb(),
+      execution,
+      runnerInstanceId: "runner",
+    });
+    await vi.waitFor(() => expect(state.release).toBeTypeOf("function"));
+
+    await expect(cancelNativeSession(execution.binding.runId, "budget hard stop"))
+      .rejects.toThrow("transport unavailable");
+    await expect(cancelNativeSession(execution.binding.runId, "retry budget stop"))
+      .resolves.toBe(true);
+    expect(state.cancel).toHaveBeenNthCalledWith(2, { reason: "retry budget stop" });
+
+    state.release?.();
+    await running;
+  });
 });

server/src/services/native-runtime/native-session-executor.ts

@@ -23,8 +23,13 @@ export async function cancelNativeSession(runId: string, reason: string): Promis
   if (!active) return false;
   if (active.cancelRequested) return true;
   active.cancelRequested = true;
-  if (active.session.cancel) await active.session.cancel({ reason });
-  else if (active.session.interrupt) await active.session.interrupt({ reason });
+  try {
+    if (active.session.cancel) await active.session.cancel({ reason });
+    else if (active.session.interrupt) await active.session.interrupt({ reason });
+  } catch (error) {
+    active.cancelRequested = false;
+    throw error;
+  }
   return true;
 }
 

server/src/services/native-runtime/paperclip-control-plane-port.test.ts

@@ -43,6 +43,7 @@ describe("PaperclipControlPlanePort conformance", () => {
   const taskIssueId = "00000000-0000-4000-8000-000000000008";
   const taskContractId = "00000000-0000-4000-8000-000000000009";
   const taskRunId = "00000000-0000-4000-8000-000000000010";
+  const taskSessionId = "00000000-0000-4000-8000-000000000018";
   const taskWorkProductId = "00000000-0000-4000-8000-000000000017";
   const workspaceFailureIssueId = "00000000-0000-4000-8000-000000000011";
   const workspaceFailureContractId = "00000000-0000-4000-8000-000000000012";
@@ -135,6 +136,7 @@ describe("PaperclipControlPlanePort conformance", () => {
       runtimeMode: "native",
       runtimeModeResolverVersion: "phase6-v1",
       runtimeModeReason: "eligible_opt_in",
+      nativeSessionId: taskSessionId,
       completionContractId: taskContractId,
       completionContractSha256: "phase6-task-contract",
       contextSnapshot: { issueId: taskIssueId },
@@ -297,7 +299,7 @@ describe("PaperclipControlPlanePort conformance", () => {
 
   it("completes one selected Paperclip task through the public package session contract", async () => {
     const identity = CONTROL_PLANE_CONFORMANCE_OPEN.identity;
-    const sessionId = "session-phase6-paperclip-task";
+    const sessionId = taskSessionId;
     const evidenceRef = `work_product:${taskWorkProductId}`;
     const taskResult = structuredClone(CONTROL_PLANE_CONFORMANCE_RESULT);
     taskResult.completionClaim.criteria[0]!.evidenceRefs = [evidenceRef];
@@ -558,12 +560,14 @@ describe("PaperclipControlPlanePort conformance", () => {
     await expect(db.select().from(nativeRunResults).where(eq(nativeRunResults.runId, workspaceFailureRunId))).resolves.toHaveLength(1);
   });
 
-  it("LIVE-01..04 rolls back status, decision, and liveness rows at every materialization failpoint", async () => {
+  it("LIVE-01..06 rolls back status, decision, and liveness rows at every materialization failpoint", async () => {
     const identity = CONTROL_PLANE_CONFORMANCE_OPEN.identity;
     const cases: Array<{
       suffix: number;
-      failpoint: "interaction_materialization" | "continuation_materialization" | "blocker_materialization" | "status_projection";
+      failpoint: "governance_materialization" | "interaction_materialization" | "continuation_materialization" | "blocker_materialization" | "recovery_materialization" | "status_projection";
       result: PrpStructuredRunResult;
+      terminalState?: "succeeded" | "failed" | "cancelled";
+      executionState?: Record<string, unknown>;
     }> = [
       {
         suffix: 20,
@@ -593,6 +597,18 @@ describe("PaperclipControlPlanePort conformance", () => {
         failpoint: "status_projection",
         result: { ...structuredClone(CONTROL_PLANE_CONFORMANCE_RESULT), reportedWorkDisposition: "needs_review" },
       },
+      {
+        suffix: 25,
+        failpoint: "recovery_materialization",
+        terminalState: "cancelled",
+        result: structuredClone(CONTROL_PLANE_CONFORMANCE_RESULT),
+      },
+      {
+        suffix: 26,
+        failpoint: "governance_materialization",
+        executionState: { status: "pending", stage: "approval" },
+        result: structuredClone(CONTROL_PLANE_CONFORMANCE_RESULT),
+      },
     ];
     for (const entry of cases) {
       const value = String(entry.suffix).padStart(12, "0");
@@ -606,6 +622,7 @@ describe("PaperclipControlPlanePort conformance", () => {
         status: "in_progress",
         assigneeAgentId: identity.agentId,
         workMode: "standard",
+        executionState: entry.executionState,
       });
       await db.insert(completionContracts).values({
         id: contractId,
@@ -634,6 +651,7 @@ describe("PaperclipControlPlanePort conformance", () => {
       });
       const terminal: PrpTerminalState = {
         ...CONTROL_PLANE_CONFORMANCE_TERMINAL,
+        runTerminalState: entry.terminalState ?? "succeeded",
         reportedWorkDisposition: entry.result.reportedWorkDisposition,
       };
       const port = new PaperclipControlPlanePort(db, {