Contain backup upload credentials and prepare provider recovery gates

Assistant: codex
Assistant-Model: gpt-6-astra
Assistant-Session: 01a06ecb-456a-71c2-b41e-0755d336e883

docs/backup-credential-recovery.md

@@ -0,0 +1,67 @@
+# WP-0029 provider recovery procedure
+
+Scope: invalidate the exposed Nextcloud upload predecessor and prove replacement
+encrypted upload and offsite recovery under CCR-2026-0004. The route is a
+Nextcloud file-drop share, not a platform-admin OpenBao credential. OpenBao
+administration does not confer Nextcloud share-owner authority.
+
+## Required owner inputs
+
+- Nextcloud share owner/account and the reviewed share-management session.
+- A replacement file-drop share/token delivered through protected custody;
+  never chat, argv, Git or State Hub. Record only custody coordinates.
+- Owner-authorized retrieval access for the encrypted offsite artifact. Do not
+  assume an upload-only file-drop token can list or download backups.
+- A protected reference to the predecessor for a bounded rejection probe, or
+  accepted provider-side revocation evidence. Do not reconstruct it into logs.
+
+## Execution order
+
+1. Confirm current OpenBao KV version and preserve every existing field. The
+   lane contains `AGE_PRIVATE_KEY`; replacing only upload fields must use CAS
+   and preserve recovery escrow. Never use a partial `kv put` that drops it.
+2. Confirm the exact Nextcloud share and all backup consumers. Besides the
+   workstation helpers, activity-core consumes `NC_WEBDAV_TOKEN` and
+   `NC_WEBDAV_URL` through ExternalSecret `actcore-backup-offsite`. Coordinate
+   its next invocation so process-cached credentials cannot produce a false
+   replacement-success claim.
+3. Create the replacement under provider authority. Verify its permissions
+   match the file-drop lane; no production delete/admin authority is needed.
+   Seed both upload fields through contained OpenBao CAS and verify delivery
+   with boolean comparisons. Keep provider secrets out of evidence.
+4. Upload a uniquely named age-encrypted artifact with the replacement. Require
+   HTTP 200/201/204; redirects and network errors are not success. Record only
+   artifact identity and safe status. A synthetic fixture can prove transport,
+   but it does not establish Forgejo application recovery.
+5. Invalidate the predecessor at the provider. Verify rejection with the old
+   credential in protected memory against the same bounded operation. A 401/403
+   or accepted provider invalidation receipt is evidence; a timeout is not.
+   Preserve existing retained backup objects and access to their encryption key.
+6. Retrieve the newly uploaded encrypted artifact through the separate owner
+   recovery path, decrypt in a protected temporary directory, and verify its
+   expected content. For Forgejo restore acceptance, use a fetched real backup
+   and the owning isolated restore procedure with repository/blob/database
+   checks. Require the local fetched artifact to exist before invoking the
+   historical infra drill: that script otherwise creates a fresh production
+   dump, which cannot prove offsite recovery.
+7. Record provider invalidation, replacement KV version, upload/download,
+   decryption, isolated restore and cleanup receipts. Mark T02 done only after
+   the agreed recovery checks pass. No receipt may contain the predecessor's
+   value, fingerprint, length or shape.
+
+## Recovery-key exposure remains separate
+
+CCR-2026-0004 records historical exposure of the age recovery key as well as
+upload credentials. Upload-share rotation alone must not clear that taint.
+Retain old age recovery material until all retained artifacts are accounted for
+and any owner-approved re-encryption is complete. Do not silently rotate or
+discard the age key as part of this upload-token repair.
+
+## Prepared transport
+
+`lib/railiance-backup-common.sh` passes curl credentials and credential-bearing
+URLs via stdin configuration, disables curlrc, suppresses backend diagnostics,
+requires HTTPS and rejects non-success responses. The Python backend streams
+the artifact, refuses redirects and suppresses credential-bearing exception
+text. These source changes are preparation; no provider rotation, upload or
+restore is claimed by their tests.

docs/forgejo-backup.md

@@ -32,6 +32,9 @@ There is no built-in credential fallback. Missing credentials stop execution
 before any cluster dump. Local encryption dry-runs skip upload authentication.
 Provider rotation and replacement upload/restore proof are tracked separately
 in `RPF-WP-0029`; removing the source default does not prove revocation.
+The concrete owner procedure is `docs/backup-credential-recovery.md`. Upload
+credentials and credential-bearing URLs are passed to curl through stdin,
+and backend errors never print those values. Redirects are refused.
 
 Decrypt: `~/.config/age/railiance-backup.key` (same key as other Railiance backups).
 

history/2026-09-05-backup-provider-recovery-preparation.md

@@ -0,0 +1,24 @@
+# WP-0029 continuation: provider gate and transport containment
+
+The user authorized continuing with backup credential invalidation and recovery.
+Routing and CCR-2026-0004 identify a Nextcloud file-drop lane plus age recovery
+escrow. No provider share-management authority or replacement custody reference
+was supplied. Requested only owner/account and custody coordinates, not values.
+
+Found and fixed credential transport defects in the retained backup helper:
+curl argv previously contained the upload credential and credential-bearing URL;
+the Python fallback could print them through an exception traceback. Curl now
+receives escaped config on stdin with curlrc disabled, both paths require HTTPS,
+redirects/non-success responses fail, and backend diagnostics are contained.
+The Python fallback streams artifact bytes. Added tests using synthetic markers.
+
+Prepared `docs/backup-credential-recovery.md`: provider invalidation, CAS-safe
+replacement preserving age escrow, activity-core delivery, encrypted upload,
+separate owner download, and isolated offsite restore acceptance. The legacy
+infra drill's missing-file fallback makes a fresh production dump; a downloaded
+artifact must be proven present before it can qualify as offsite recovery.
+
+No provider change, credential read/provisioning, upload or restore was performed.
+The attempted metadata-only taint query lacked caller authentication. Historical
+CCR recovery-key exposure remains recorded; no current taint state or rotation
+is inferred. T02 remains waiting on the provider inputs and live proof.

lib/railiance-backup-common.sh

@@ -12,8 +12,8 @@ railiance_backup_load_openbao_lane() {
   fi
   command -v bao >/dev/null 2>&1 || return 0
   bao kv metadata get "${RAILIANCE_BACKUP_BAO_PATH}" >/dev/null 2>&1 || return 0
-  : "${RAILIANCE_BACKUP_NC_TOKEN:=$(bao kv get -field=NC_WEBDAV_TOKEN "${RAILIANCE_BACKUP_BAO_PATH}")}"
+  : "${RAILIANCE_BACKUP_NC_TOKEN:=$(bao kv get -field=NC_WEBDAV_TOKEN "${RAILIANCE_BACKUP_BAO_PATH}" 2>/dev/null)}"
-  : "${RAILIANCE_BACKUP_NC_WEBDAV_URL:=$(bao kv get -field=NC_WEBDAV_URL "${RAILIANCE_BACKUP_BAO_PATH}")}"
+  : "${RAILIANCE_BACKUP_NC_WEBDAV_URL:=$(bao kv get -field=NC_WEBDAV_URL "${RAILIANCE_BACKUP_BAO_PATH}" 2>/dev/null)}"
 }
 
 railiance_backup_require_openbao_lane() {
@@ -51,27 +51,54 @@ railiance_backup_require_tools() {
 railiance_backup_nc_upload() {
   local file="$1" remote_name="$2"
   local dest="${RAILIANCE_BACKUP_NC_WEBDAV_URL%/}/${RAILIANCE_BACKUP_NC_PREFIX}/${remote_name}"
+  if [[ "$dest" != https://* || "$dest" == *$'\n'* || "$dest" == *$'\r'* ||
+        "${RAILIANCE_BACKUP_NC_TOKEN}" == *$'\n'* || "${RAILIANCE_BACKUP_NC_TOKEN}" == *$'\r'* ||
+        "$file" == *$'\n'* || "$file" == *$'\r'* ]]; then
+    echo "ERROR: invalid backup upload input" >&2
+    return 1
+  fi
   if command -v curl >/dev/null 2>&1; then
-    curl -sf -u "${RAILIANCE_BACKUP_NC_TOKEN}:" -T "$file" "$dest" \
+    # Credentials and credential-bearing URLs travel over stdin, never argv.
-      || { echo "ERROR: Nextcloud upload failed: ${remote_name}" >&2; return 1; }
+    local config_url config_token config_file status
+    config_url="${dest//\\/\\\\}"; config_url="${config_url//\"/\\\"}"
+    config_token="${RAILIANCE_BACKUP_NC_TOKEN//\\/\\\\}"; config_token="${config_token//\"/\\\"}"
+    config_file="${file//\\/\\\\}"; config_file="${config_file//\"/\\\"}"
+    status="$(printf 'url = "%s"\nuser = "%s:"\nupload-file = "%s"\n' \
+      "$config_url" "$config_token" "$config_file" | \
+      curl --disable --silent --fail --proto '=https' --max-time 600 \
+        --output /dev/null --write-out '%{http_code}' --config - 2>/dev/null)" \
+      || { echo "ERROR: Nextcloud upload failed" >&2; return 1; }
+    case "$status" in
+      200|201|204) ;;
+      *) echo "ERROR: Nextcloud upload failed" >&2; return 1 ;;
+    esac
     return 0
   fi
   # Python fallback (worker image)
   RAILIANCE_BACKUP_NC_TOKEN="${RAILIANCE_BACKUP_NC_TOKEN}" \
   RAILIANCE_BACKUP_UPLOAD_URL="${dest}" \
   RAILIANCE_BACKUP_UPLOAD_FILE="${file}" \
-  python3 - <<'PY' || { echo "ERROR: Nextcloud upload failed: ${remote_name}" >&2; return 1; }
+  python3 - <<'PY' || { echo "ERROR: Nextcloud upload failed" >&2; return 1; }
-import os, urllib.request, base64
+import os, urllib.request, urllib.error, base64, sys
 url = os.environ["RAILIANCE_BACKUP_UPLOAD_URL"]
 path = os.environ["RAILIANCE_BACKUP_UPLOAD_FILE"]
 token = os.environ["RAILIANCE_BACKUP_NC_TOKEN"]
-data = open(path, "rb").read()
+class NoRedirect(urllib.request.HTTPRedirectHandler):
-req = urllib.request.Request(url, data=data, method="PUT")
+    def redirect_request(self, req, fp, code, msg, headers, newurl):
-req.add_header("Authorization", "Basic " + base64.b64encode(f"{token}:".encode()).decode())
+        raise urllib.error.HTTPError(req.full_url, code, 'redirect refused', headers, fp)
-req.add_header("Content-Type", "application/octet-stream")
+try:
-with urllib.request.urlopen(req, timeout=600) as resp:
+    with open(path, 'rb') as source:
-    if resp.status not in (200, 201, 204):
+        chunks = iter(lambda: source.read(1024 * 1024), b'')
-        raise SystemExit(f"HTTP {resp.status}")
+        req = urllib.request.Request(url, data=chunks, method="PUT")
+        req.add_header("Authorization", "Basic " + base64.b64encode(f"{token}:".encode()).decode())
+        req.add_header("Content-Type", "application/octet-stream")
+        req.add_header("Content-Length", str(os.fstat(source.fileno()).st_size))
+        with urllib.request.build_opener(NoRedirect()).open(req, timeout=600) as resp:
+            if resp.status not in (200, 201, 204):
+                sys.exit(1)
+except Exception:
+    # URLs may themselves contain credentials: never emit exception text.
+    sys.exit(1)
 PY
 }
 
@@ -91,4 +118,4 @@ railiance_backup_record_success() {
   mkdir -p "$stamp_dir"
   echo "$ts" > "${stamp_dir}/.last-success"
   echo "$ts" >> "${stamp_dir}/success-log"
-}
+}

tests/test_backup_credentials.py

@@ -33,6 +33,78 @@ railiance_backup_require_openbao_lane
         self.assertEqual(result.returncode, 0)
         self.assertEqual(result.stdout + result.stderr, "")
 
+    def test_curl_receives_credentials_only_on_stdin(self):
+        result = self.run_shell('''
+RAILIANCE_BACKUP_NC_TOKEN=fixture-secret-marker
+RAILIANCE_BACKUP_NC_WEBDAV_URL=https://example.invalid/fixture-secret-marker
+curl() {
+  [[ "$*" != *fixture-secret-marker* ]] || return 9
+  local config
+  config="$(cat)"
+  [[ "$config" == *'user = "fixture-secret-marker:"'* ]] || return 8
+  [[ "$config" == *'url = "https://example.invalid/fixture-secret-marker/forgejo/test.age"'* ]] || return 7
+  printf 201
+}
+railiance_backup_nc_upload /tmp/test.age test.age
+''')
+        self.assertEqual(result.returncode, 0, result.stderr)
+        self.assertEqual(result.stdout + result.stderr, '')
+
+    def test_redirect_and_backend_diagnostics_are_not_success_or_output(self):
+        for status in ['302', '401', '403', '500']:
+            result = self.run_shell('''
+RAILIANCE_BACKUP_NC_TOKEN=fixture-secret-marker
+RAILIANCE_BACKUP_NC_WEBDAV_URL=https://example.invalid/fixture-secret-marker
+curl() { cat >/dev/null; echo fixture-secret-marker >&2; printf '%s'; }
+railiance_backup_nc_upload /tmp/test.age test.age
+''' % status)
+            self.assertNotEqual(result.returncode, 0)
+            self.assertEqual(result.stdout, '')
+            self.assertEqual(result.stderr, 'ERROR: Nextcloud upload failed\n')
+
+    def test_config_line_injection_fails_before_curl(self):
+        result = self.run_shell('''
+RAILIANCE_BACKUP_NC_TOKEN=$'fixture\\noutput=/tmp/unwanted'
+RAILIANCE_BACKUP_NC_WEBDAV_URL=https://example.invalid/upload
+curl() { echo backend-was-called; return 1; }
+railiance_backup_nc_upload /tmp/test.age test.age
+''')
+        self.assertNotEqual(result.returncode, 0)
+        self.assertEqual(result.stdout, '')
+        self.assertEqual(result.stderr, 'ERROR: invalid backup upload input\n')
+
+    def test_plain_http_is_rejected_before_backend(self):
+        result = self.run_shell('''
+RAILIANCE_BACKUP_NC_TOKEN=fixture
+RAILIANCE_BACKUP_NC_WEBDAV_URL=http://example.invalid/upload
+curl() { echo backend-was-called; return 1; }
+railiance_backup_nc_upload /tmp/test.age test.age
+''')
+        self.assertNotEqual(result.returncode, 0)
+        self.assertEqual(result.stdout, '')
+
+    def test_python_backend_exception_cannot_disclose_credential_url(self):
+        result = self.run_shell('''
+RAILIANCE_BACKUP_NC_TOKEN=fixture-secret-marker
+RAILIANCE_BACKUP_NC_WEBDAV_URL=https://example.invalid/fixture-secret-marker
+command() {
+  if [[ "$*" == '-v curl' ]]; then return 1; fi
+  builtin command "$@"
+}
+python3() {
+ /usr/bin/python3 -c 'import sys,urllib.request
+class FailedOpener:
+ def open(self, req, **kwargs):
+  raise RuntimeError(req.full_url)
+urllib.request.build_opener=lambda *args: FailedOpener()
+exec(sys.stdin.read())'
+}
+railiance_backup_nc_upload /dev/null test.age
+''')
+        self.assertNotEqual(result.returncode, 0)
+        self.assertEqual(result.stdout, '')
+        self.assertEqual(result.stderr, 'ERROR: Nextcloud upload failed\n')
+
     def test_openbao_lane_supplies_missing_credentials(self):
         result = self.run_shell('''
 bao() {

workplans/RPF-WP-0029-backup-credential-default-removal.md

@@ -27,6 +27,11 @@ state_hub_task_id: "4b5aefdb-a746-54f9-ba29-ebb840e7848d"
 
 Removed the literal upload credential default from tools/cmd/forgejo-backup. Missing governed input now fails before cluster operations with a value-free diagnostic; encryption dry-runs skip upload authentication. Offline tests prove missing-input denial, explicit input, and mocked OpenBao resolution.
 
+2026-09-05 continuation: removed credentials and credential-bearing URLs from
+curl argv; curl config travels through stdin with curlrc disabled. Both upload
+backends suppress credential-bearing diagnostics, require HTTPS and reject
+redirects/non-success status. Added transport containment and failure tests.
+
 ## Invalidate predecessor and prove replacement recovery
 
 ```task
@@ -38,6 +43,13 @@ state_hub_task_id: "b3f3402f-890b-5781-9b3e-1c9c0d28cea8"
 
 Provider-side invalidation and replacement custody need the attended provider owner and CCR-2026-0004 lifecycle procedure. Record only non-secret invalidation, encrypted upload and restore receipts. No provider authority or replacement receipt was available; source removal alone does not close the reported exposure. Never record the predecessor value, fingerprint, length, or shape.
 
+Prepared owner execution procedure: `docs/backup-credential-recovery.md`.
+Awaiting the Nextcloud share owner/account and replacement custody coordinates;
+OpenBao authority is not provider share-management authority. Activity-core is
+also a consumer of this upload lane. Preserve AGE_PRIVATE_KEY and historical
+exposure evidence; upload-token rotation cannot clear recovery-key taint.
+No provider mutation, backup upload or restore has been run in this continuation.
+
 ## Portfolio review — 2026-09-05
 
 INTENT binding: dependable backup custody and recovery. Retain T02 until the