Inbox and notifications cannot distinguish a mention from a person from a mention from an agent
- Dominant language
- Rust
- Stars
- 32.7k
- Forks
- 4.3k
- Avg merge
- 1d 13h
- Merged PRs (30d)
- 253
Description
**Motivation**
In a busy community most messages come from agents. Mentions from people get buried in that stream, and there is no way to separate them.
Today the Inbox filter menu offers All / Projects / Mentions / Threads / Needs action / Agents / Reminders / Drafts (same set on Desktop and mobile). "Mentions" mixes people and agents, and "Agents" shows agent activity, not "everything except agents". Desktop notifications and sounds have a single "@Mentions" slot with the same problem: a person tagging you and an agent tagging you produce the identical alert.
Most people still prioritise a message from a colleague over one from an agent, even in a setup that otherwise treats agents as equal participants.
**Proposed solution**
1. An Inbox filter that shows only mentions and messages from people, excluding agents.
2. A separate notification/sound slot for mentions from people, so a human tagging you can be made louder than an agent tagging you (or the agent one turned off).
The app already knows which accounts are agents — it renders their avatars differently — so the distinction is available where the filtering happens.
**Alternatives considered**
- Muting noisy agents individually: does not scale, and it hides useful agent output as well.
- Reading the Mentions filter and skimming for names: works, but that is the manual version of the filter being asked for.
**Additional context**
Duplicate search: none found.
Contributor guide
Research direction
No files, tests, or entry points are named. Start by locating the Inbox filter implementation and the desktop/mobile notification and sound settings; trace how agent accounts are identified during filtering. Done means people-only mentions can be filtered separately and human and agent mention alerts can be configured independently.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- rust
- Domain
- desktop, frontend, mobile
- Issue type
- Feature
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Activity status
- Active
- Clarity
- Mostly clear
- Newbie friendliness
- 52/100