fix(chat): the operator was missing from the roster they host

Found on two live machines within minutes of deploying: every message from the
person running the node arrived at every other member under "this account is
using a key you have not seen before".

An operator's authority is node-wide and is recorded in `members` with an
**empty** group_id — `is_authorized` has always said so, in a clause written for
exactly that. `group_devices` spelled the rule out a second time as
`WHERE m.group_id = ?`, which excludes them. So the operator was absent from the
roster relayed to members, no chain could reach their device key, and Tier 2
reported the most ordinary event there is — the operator talking in their own
group — as a key substitution.

A notice that fires on normal use is worse than no notice: it is the one people
learn to dismiss, and §4.8 budgets exactly one for the whole feature. That makes
this a defect in the property, not only in a query.

The clause now lives once, as `_MEMBER_OF_GROUP`, shared by both callers so they
cannot drift again. `DISTINCT` because an operator who is also an explicit member
of the group matches both halves of it. `get_member` is untouched: it is a raw
lookup and its callers already fall back to `get_member("", user_id)` themselves.

Three tests, and the first fails against the old query with the reported
symptom: the operator appears in the roster of a group they host and
`is_authorized` agrees; an operator who is also a member is listed once; a
revoked one comes back through neither.

No stored state to clean up — nothing was written to a client's pins when the
account was missing, so the notice stops as soon as the node serves the roster
correctly.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01TZZxYjz8YeWRz13xDi8LJr

1 files changed, 72 insertions, 0 deletions

diff --git a/packages/meshbay-node/tests/test_group_roster.py b/packages/meshbay-node/tests/test_group_roster.py
index d7ca7dc..cb9828c 100644
--- a/packages/meshbay-node/tests/test_group_roster.py
+++ b/packages/meshbay-node/tests/test_group_roster.py
@@ -240,3 +240,75 @@ async def test_a_substituted_key_carries_no_evidence(tmp_path, roster):
     assert devices[pk_evil]["add_sig"] == "", (
         "a fabricated device cannot come with a countersignature — if this ever "
         "holds evidence, the node has been handed a way to mint trust")
+
+
+# ── the operator is a member of every group this node hosts ──────────────────
+
+async def test_the_operator_is_in_the_roster_of_a_group_they_host(tmp_path,
+                                                                  roster):
+    """
+    Found on two live machines, and it is the shape this file exists to stop.
+
+    An operator's authority is node-wide and is stored with an **empty**
+    group_id (`is_authorized` says so, and has always said so). `group_devices`
+    wrote that rule out a second time as `WHERE m.group_id = ?`, which excludes
+    them — so the person running the node was absent from the roster relayed to
+    everyone else, their device key could be vouched for by nobody, and every
+    single message they sent arrived under "this account is using a key you
+    have not seen before".
+
+    A notice that fires on the most ordinary event there is — the operator
+    talking in their own group — is worse than no notice, because it is the one
+    people learn to dismiss. Both queries now share `_MEMBER_OF_GROUP`.
+    """
+    _sk_op, pk_op, px_op = _keys()
+    _sk_m, pk_m, px_m = _keys()
+    # The operator pairs node-wide: group_id is empty, exactly as
+    # `pairOperator` sends it and `_pin_and_admit` records it.
+    await roster.pin_identity("toto", "toto", pk_op, px_op, via="code")
+    await roster.set_member(group_id="", user_id="toto", role="operator",
+                            status="active", approved_by="self")
+    # An invited member of one group this node hosts.
+    await roster.pin_identity("cbesson", "cbesson", pk_m, px_m, via="invite")
+    await roster.set_member(group_id=GROUP, user_id="cbesson",
+                            role=ROLE_MEMBER, status="active",
+                            approved_by="toto")
+
+    users = {d["user_id"] for d in await roster.group_devices(GROUP)}
+    assert users == {"toto", "cbesson"}, (
+        "the operator must appear in the roster of a group they host — "
+        "otherwise every message they send reads as an unknown key")
+
+    # And the two rules genuinely agree, rather than happening to agree here.
+    assert await roster.is_authorized(GROUP, "toto") is True
+
+
+async def test_an_operator_who_is_also_a_member_appears_once(tmp_path, roster):
+    """
+    Both halves of the clause match such a person. Listed twice, a client would
+    see the same key arrive as two devices — harmless today, and exactly the
+    kind of thing that grows teeth later.
+    """
+    _sk_op, pk_op, px_op = _keys()
+    await roster.pin_identity("toto", "toto", pk_op, px_op, via="code")
+    await roster.set_member(group_id="", user_id="toto", role="operator",
+                            status="active", approved_by="self")
+    await roster.set_member(group_id=GROUP, user_id="toto", role=ROLE_MEMBER,
+                            status="active", approved_by="self")
+
+    devices = await roster.group_devices(GROUP)
+    assert [d["pk_ed25519"] for d in devices] == [pk_op]
+
+
+async def test_an_operator_of_another_node_is_not_invented(tmp_path, roster):
+    """
+    The clause admits an operator, not anyone with an empty group_id. A
+    revoked or suspended one must not come back through it.
+    """
+    _sk_op, pk_op, px_op = _keys()
+    await roster.pin_identity("toto", "toto", pk_op, px_op, via="code")
+    await roster.set_member(group_id="", user_id="toto", role="operator",
+                            status="revoked", approved_by="self")
+
+    assert await roster.group_devices(GROUP) == []
+    assert await roster.is_authorized(GROUP, "toto") is False