fix(test): preserve sandbox refusal provenance

docs/scratchpads/1029-rm-02-gate-registry.md

@@ -61,7 +61,7 @@ The queue guard's `get_state_from_status_json` runs `python3 - <<'PY'` while pro
 - Initial PR pipeline #2177 exposed Woodpecker's shallow boundary: the activation parent object was present but marked shallow, so `merge-base --is-ancestor` correctly refused to infer ancestry. The unconditional gate step now unshallows before ancestry/provenance checks; its wiring test was observed RED before the CI fix.
 - Pipeline #2178 then proved the unprivileged Docker runner cannot establish Bubblewrap namespaces. A privileged experiment remained uncommitted and was rejected after Codex correctly rated it CRITICAL: PR-controlled code executes before an in-repository sandbox and could directly use the granted capability.
 - `mos-remediation` and `rev-974` independently ruled Option C. RM02-REQ-10 now retains its original text, restatement, and reason: PR CI verifies only the current tree, unprivileged and fail-closed; isolated own-tree replay is deferred to RM-60/#1031's external pre-execution authority, cross-referenced with RM-59. Future protected post-merge replay is detection with quarantine/revert, never pre-merge prevention.
-- RED-first boundary test proved the old path executed an inert intermediate verifier. The revised path states adjacent `DOES`/`DOES NOT` claims, validates historical manifest provenance without executing it, and infers no replay success. Direct sandbox tests remain hard-fail; unprivileged CI asserts terminal refusal instead of treating replay as success. Pipeline #2179 showed this runner reports namespace denial as `spawnSync bwrap` with `error.code=EPERM`, `status=null`, and no stderr; the refusal detector now recognizes only that Bubblewrap-provenance hard-fail form and rejects unrelated command EPERM results.
+- RED-first boundary test proved the old path executed an inert intermediate verifier. The revised path states adjacent `DOES`/`DOES NOT` claims, validates historical manifest provenance without executing it, and infers no replay success. Direct sandbox tests remain hard-fail; unprivileged CI asserts terminal refusal instead of treating replay as success. Pipelines #2179/#2180/#2181 exposed two runner refusal forms: namespace denial as `spawnSync bwrap` with `error.code=EPERM`, and a test image without Bubblewrap as `error.code=ENOENT`. The replay diagnostic now preserves spawn errors; the refusal detector recognizes only exact `spawnSync bwrap` provenance for `EPERM`/`EACCES`/`ENOENT`, plus Bubblewrap's known namespace-refusal text. Focused negative assertions reject both unrelated `spawnSync git EPERM` and verifier output that merely says `bwrap ENOENT`.
 - Option C security review reported no findings. Code review rejected an initial unrelated typecheck binding for the new security criterion. It was replaced with a dedicated registered `privileged-pr-gate` case: the fixture injects a privilege key into the gate step, the wiring control rejects it for that exact reason, and `gate:verify` observes the boundary negative control. Follow-up hardening uses a closed exact gate-step construction, rejects privilege across the entire pipeline, rejects non-canonical/merged YAML keys, and pins the unrestricted PR/main trigger block; quoted/escaped/alias/merge/duplicate/filter bypass tests pass. Final Codex code review approved with no findings.
 
 ## Documentation checklist

scripts/gate-history.mjs

@@ -142,7 +142,7 @@ export async function replayCommit(root, commit) {
       if (install.status !== 0 || install.error || install.signal) {
         return {
           ...install,
-          stderr: `historical frozen dependency install failed: ${install.stderr || install.stdout || ''}`,
+          stderr: `historical frozen dependency install failed: ${install.error?.message || install.stderr || install.stdout || ''}`,
         };
       }
       const authoritativeChanges = await authoritativeTreeChanges(

scripts/gate-history.test.mjs

@@ -17,10 +17,11 @@ const fixtureRoot = path.join(process.cwd(), '.mosaic-test-work', `gate-history-
 function sandboxUnavailable(result) {
   const detail = `${result.stdout ?? ''}${result.stderr ?? ''}${result.error?.message ?? ''}`;
   const bubblewrapSpawnDenied =
-    result.error?.code === 'EPERM' && /spawnSync bwrap/i.test(result.error?.message ?? '');
+    ['EPERM', 'EACCES', 'ENOENT'].includes(result.error?.code) &&
+    /spawnSync bwrap/i.test(result.error?.message ?? '');
   if (
     !bubblewrapSpawnDenied &&
-    !/bwrap.*(?:EPERM|Operation not permitted|Creating new namespace failed)/i.test(detail)
+    !/bwrap:.*(?:Operation not permitted|Creating new namespace failed)/i.test(detail)
   ) {
     return false;
   }
@@ -50,13 +51,15 @@ test.after(async () => {
 });
 
 test('sandbox refusal classification requires Bubblewrap provenance', () => {
-  assert.equal(
+  for (const code of ['EPERM', 'EACCES', 'ENOENT']) {
-    sandboxUnavailable({
+    assert.equal(
-      status: null,
+      sandboxUnavailable({
-      error: { code: 'EPERM', message: 'spawnSync bwrap EPERM' },
+        status: null,
-    }),
+        error: { code, message: `spawnSync bwrap ${code}` },
-    true,
+      }),
-  );
+      true,
+    );
+  }
   assert.equal(
     sandboxUnavailable({
       status: null,
@@ -64,6 +67,10 @@ test('sandbox refusal classification requires Bubblewrap provenance', () => {
     }),
     false,
   );
+  assert.equal(
+    sandboxUnavailable({ status: 1, stderr: 'historical verifier said bwrap ENOENT' }),
+    false,
+  );
 });
 
 test('prospective history reads each commit own manifest rather than the current tree', async () => {