fix: a member can no longer lock a node, crash it with a link, or stop hub cleanup

- node: only a wrong code counts towards the join lock, now per account
  (5) as well as node-wide (20), and it is consulted only when a code is
  tried. Every member reconnecting gets the group key through join_request,
  so a lock checked before recognition let one member refuse it to everyone.
- node: link previews read the body as a stream and stop at the cap,
  counted on decoded bytes; a declared oversized image is not read; 15 s
  total deadline; image decoding off the loop. `client.get` had buffered
  the whole (decompressed) response before the caps looked at it.
- hub: the daily purge of never-verified accounts detaches their IP-log
  rows (keeping the name) and clears every other reference first, and each
  cleanup step runs on its own. On PostgreSQL the bare DELETE violated the
  ip_logs foreign key and stopped every purge behind it for good.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

1 files changed, 55 insertions, 19 deletions

diff --git a/packages/meshbay-node/src/meshbay_node/transport/webrtc/admission.py b/packages/meshbay-node/src/meshbay_node/transport/webrtc/admission.py
index e678a13..f94cade 100644
--- a/packages/meshbay-node/src/meshbay_node/transport/webrtc/admission.py
+++ b/packages/meshbay-node/src/meshbay_node/transport/webrtc/admission.py
@@ -40,9 +40,18 @@ _INVITE_ID_RE = re.compile(r"[0-9a-f]{32}")
 MAX_JOIN_ATTEMPTS = 5
 # Per-connection limits alone would not bind an attacker who can open connections
 # at will — and the adversary who can mint tokens for any account is the hub. So
-# failed pairings are also counted node-wide over a window.
+# wrong codes are also counted per account and node-wide over a window.
+#
+# Only a *wrong code* counts, and the lock is consulted only when a code is about
+# to be tried. Every member reconnecting obtains the group key through this very
+# message — no per-member bundle is stored — so a lock checked at the top of it,
+# fed by ordinary refusals such as `code_required`, let any one member refuse the
+# key to everybody on the node for as long as they kept failing. A device the
+# node already pinned, joining without a code, is never subject to it.
 MAX_JOIN_FAILURES_WINDOW = 20
+MAX_JOIN_FAILURES_PER_ACCOUNT = 5
 JOIN_FAILURE_WINDOW = 600  # seconds
+_JOIN_FAILURE_ACCOUNTS_TRACKED = 1000
 
 
 class AdmissionMixin:
@@ -124,15 +133,11 @@ class AdmissionMixin:
 
     # ── Pairing and join (H3, M3) ────────────────────────────────────────────
 
-    def _join_refuse(self, reason: str, audit_detail: str = "") -> None:
+    def _join_refuse(self, reason: str, audit_detail: str = "", *,
+                     wrong_code: bool = False) -> None:
         self._join_attempts += 1
-        # Node-wide window, shared across connections: reconnecting must not reset
-        # the budget.
-        now = time.time()
-        failures = [t for t in self._ctx.get("join_failures", [])
-                    if now - t < JOIN_FAILURE_WINDOW]
-        failures.append(now)
-        self._ctx["join_failures"] = failures
+        if wrong_code:
+            self._record_wrong_code()
         self._audit_join("join_refused", audit_detail or reason)
         self._send({
             "type": MNP.JOIN_RESULT,
@@ -141,6 +146,41 @@ class AdmissionMixin:
             "reason": reason,
         })
 
+    def _record_wrong_code(self) -> None:
+        """Count one wrong code, per account and node-wide, across connections:
+        reconnecting must not reset either budget."""
+        now = time.time()
+        failures = [t for t in self._ctx.get("join_failures", [])
+                    if now - t < JOIN_FAILURE_WINDOW]
+        failures.append(now)
+        self._ctx["join_failures"] = failures
+
+        per_account: dict = self._ctx.setdefault("join_failures_by_account", {})
+        if len(per_account) > _JOIN_FAILURE_ACCOUNTS_TRACKED:
+            for uid, times in list(per_account.items()):
+                if not times or now - times[-1] >= JOIN_FAILURE_WINDOW:
+                    per_account.pop(uid, None)
+        uid = self._user_id or getattr(self, "_pending_sub", "")
+        mine = [t for t in per_account.get(uid, ()) if now - t < JOIN_FAILURE_WINDOW]
+        mine.append(now)
+        per_account[uid] = mine
+
+    def _code_guessing_locked(self, user_id: str) -> bool:
+        """Whether a code may be tried now. Refuses, and says so, when not."""
+        now = time.time()
+        recent = [t for t in self._ctx.get("join_failures", [])
+                  if now - t < JOIN_FAILURE_WINDOW]
+        mine = [t for t in (self._ctx.get("join_failures_by_account") or {})
+                .get(user_id, ()) if now - t < JOIN_FAILURE_WINDOW]
+        if len(mine) >= MAX_JOIN_FAILURES_PER_ACCOUNT:
+            self._audit_join("join_throttled", f"{len(mine)} wrong codes by this account")
+        elif len(recent) >= MAX_JOIN_FAILURES_WINDOW:
+            self._audit_join("join_throttled", f"{len(recent)} wrong codes on this node")
+        else:
+            return False
+        self._send({"type": "error", "detail": "Pairing temporarily locked"})
+        return True
+
     def _audit_join(self, event: str, detail: str) -> None:
         audit = self._ctx.get("audit_store")
         if not audit:
@@ -177,14 +217,6 @@ class AdmissionMixin:
             self._send({"type": "error", "detail": "Too many attempts"})
             return
 
-        now = time.time()
-        recent = [t for t in self._ctx.get("join_failures", [])
-                  if now - t < JOIN_FAILURE_WINDOW]
-        if len(recent) >= MAX_JOIN_FAILURES_WINDOW:
-            self._audit_join("join_throttled", f"{len(recent)} failures in window")
-            self._send({"type": "error", "detail": "Pairing temporarily locked"})
-            return
-
         user_id = self._user_id or getattr(self, "_pending_sub", "")
         username = self._username or getattr(self, "_pending_username", "")
         if not user_id:
@@ -295,9 +327,11 @@ class AdmissionMixin:
                 if not code:
                     self._join_refuse("code_required")
                     return
+                if self._code_guessing_locked(user_id):
+                    return
                 invite = await roster.consume_invite(code, user_id, session_group)
                 if not invite:
-                    self._join_refuse("code_invalid")
+                    self._join_refuse("code_invalid", wrong_code=True)
                     return
                 await roster.set_member(
                     group_id=invite["group_id"], user_id=user_id,
@@ -337,9 +371,11 @@ class AdmissionMixin:
             self._join_refuse("code_required")
             return
 
+        if self._code_guessing_locked(user_id):
+            return
         invite = await roster.consume_invite(code, user_id, session_group)
         if not invite:
-            self._join_refuse("code_invalid")
+            self._join_refuse("code_invalid", wrong_code=True)
             return
 
         await self._pin_and_admit(