Skip to content

Commit ec508fd

Browse files
authored
Add public minutes for June 2025 MGM (#137)
1 parent 42a67b8 commit ec508fd

File tree

1 file changed

+64
-0
lines changed

1 file changed

+64
-0
lines changed
Lines changed: 64 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,64 @@
1+
---
2+
title: Monthly General Meeting, June 2025
3+
author: She
4+
excerpt: Internal policy about when to use the spamfilter,
5+
spamfilter upgrades, and guidelines on "security" status for staff.
6+
---
7+
8+
## Propositions and Motions
9+
10+
There were no propositions or motions for this meeting.
11+
12+
## Other Questions
13+
14+
### Spamfilter usage and features
15+
16+
Libera.Chat has a spam filter module that is loaded on all IRC servers.
17+
It is possible to opt out of spam filtration with channel mode `u`
18+
for messages sent to channels and user mode `u` for messages received as PMs.
19+
Additionally, the spam filter can only inform staff of which user tripped the
20+
filter, not the contents or target of the message that was filtered.
21+
22+
Spam filtration inherently comes with a risk of false positives, especially
23+
given that users might quote spam after it gets added to the filter.
24+
Therefore, it was generally understood that items should only be added to the
25+
spam filter when nothing else is effective at mitigation, and only with
26+
strong informal agreement from other members of staff.
27+
However, prior to this meeting, there was no explicit internal policy for
28+
when to add entries to the spamfilter. This meeting agreed to create an
29+
internal policy document for spam filter usage.
30+
31+
This meeting also discussed the possibility of having spam filter
32+
patterns that cannot be opted out of. Following internal discussions outside
33+
of general meetings, it was agreed that a `superdrop` flag type should be
34+
implemented that causes messages to be discarded even when the target
35+
is `+u`. It was decided that `+u` should still allow the sender to not trigger
36+
any behaviour that notifies staff of the filtered message.
37+
As before, private messages to staff would still bypass the filter.
38+
39+
It was also discussed how we should handle client-side exploits in the context
40+
of spam filtration, such as denial of service attacks exploiting overeager
41+
antivirus software or remote code execution on scriptable clients.
42+
It was agreed that both classes of exploits should be filtered, but only
43+
remote code execution exploit mitigation should use the `superdrop` flag.
44+
45+
### Guidelines around the "security" status for staff
46+
47+
There exists an internal "security" status in our service management system
48+
that can be applied by the operations team in case of a security breach.
49+
When applied to a staffer, they are stripped of most of their privileges
50+
except for their reduced IRC connection limits, their email account,
51+
and access to a shell account on an internal server.
52+
53+
There hasn't been a need for this status to be applied so far, but
54+
an item was added to the agenda to draft internal policy for its use.
55+
It was quickly agreed that no explicit internal policy for using this status
56+
is necessary at this time. It was, however, suggested that if the operations
57+
team uses this status, they must notify the board.
58+
59+
It was also mentioned that applying the status as-is could potentially
60+
violate bylaws, as members with the security status are still members of the
61+
organization but would not have access to the mailing list through which
62+
general meeting invites are sent. It was agreed that those under this status
63+
should be able to keep their access to the mailing list, as the operations
64+
team can simply remove their email access if it is being abused.

0 commit comments

Comments
 (0)