mirror of
https://github.com/paperclipai/paperclip.git
synced 2026-10-06 21:05:21 +02:00
DottaandPaperclip
5282cabde8
## Thinking Path
> - Paperclip is the open source app people use to manage AI agents for
work.
> - Experimental Agent Chat gives each person a persistent conversation
with each agent.
> - The sidebar only showed starred or previously visited agents.
> - A person could not start a conversation with an agent absent from
that list.
> - This pull request adds a searchable agent picker and a default
sidebar entry.
> - People can now reach every company agent while keeping their
frequent conversations close.
## Linked Issues or Issue Description
Refs #13283, which introduced the existing task-backed agent chat
surface. This change adds discovery to that surface after maintainer
review of the Storybook design.
**What happened?**
With Agent Chat enabled, a person with no starred or recent agents had
no agent shortcut. The sidebar also lacked a direct way to start a chat
with another agent.
**Expected behavior**
Show the first-created company agent by default. Let the person search
all company agents and open a persistent conversation.
**Steps to reproduce**
1. Enable Agent Chat in a company with agents.
2. Use an account with no starred agents or recent chats.
3. Look for a chat entry in the sidebar, or try to chat with an agent
absent from its shortcuts.
**Paperclip version or commit**
The existing behavior is present on master at 728f7185f.
## What Changed
- Add a compact picker that searches every company agent by name or
role. Include loading, retry, empty, and no-match states.
- Keep the first-created agent in Chats alongside stars and up to four
other recent conversations.
- Align the compose button with stars. Show it on hover or keyboard
focus, and keep it visible on touch devices.
- Use the existing company-aware chat route. Close the mobile sidebar on
selection and reset the picker when the company or user changes.
- Connect the approved Storybook scenarios to the production sidebar and
picker. Update product and design-guide documentation.
## Verification
- Ten focused tests pass for ordering, duplicate names, role search,
keyboard selection, search reset, retry states, full-roster navigation,
mobile dismissal, and company/user changes.
- The agent-chat sidebar end-to-end test passes against an isolated
PostgreSQL container. It verifies the first-use default, picker
navigation, lazy creation, stars, draft preservation, and reopening
existing history. The old four-row and See all link expectations now
match the approved behavior.
- Browser review of the production components in Storybook: open the
sidebar picker, search for an agent absent from the shortcuts, select
with Enter, and send a message through the existing composer.
- `pnpm build`, `pnpm -r typecheck`, `pnpm check:token-gates`, and `pnpm
build-storybook` pass on the rebased branch. The full local test run
reached 12,020 passing server tests, then stopped because an unrelated
environment-custom-images suite could not allocate PostgreSQL shared
memory on this Mac. All corresponding CI test shards pass on isolated
hosts.
- All CI checks pass on commit
`c324568324a6a8b65adfc6447ef19bcc710c61df`. Greptile reviewed that
commit at 5/5 with no open findings.
- Reviewer path: enable Agent Chat, open the Chats compose button,
search by name or role, and select an agent. Existing conversation
history must remain available.
## Risks
- Sidebar ordering changes: the earliest-created agent remains visible
even without a visit or star.
- The hover control remains available to keyboard and touch users.
- This change uses the existing feature gate, APIs, membership
preferences, and persistent conversation behavior. It adds no migration
or server contract.
- Storybook uses fixture responses. Its first-use landing page is a
review fixture and is not added as a product route.
## Model Used
OpenAI GPT-6 Astra (`gpt-6-astra`) through Codex, with reasoning, code
execution, and browser tools. The session does not expose its context
window size.
## Checklist
- [x] I have included a thinking path that traces from project context
to this change
- [x] I have specified the model used (with version and capability
details)
- [x] I have checked ROADMAP.md and confirmed this PR does not duplicate
planned core work
- [x] I have searched GitHub for duplicate or related PRs and linked
them above
- [x] I have either (a) linked existing issues with `Fixes: #` / `Closes
#` / `Refs #` OR (b) described the issue in-PR following the relevant
issue template
- [x] I have not referenced internal/instance-local Paperclip issues or
links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip`
URLs)
- [x] My branch name describes the change (e.g. `docs/...`, `fix/...`)
and contains no internal Paperclip ticket id or instance-derived details
- [x] I have run tests locally and they pass
- [x] I have added or updated tests where applicable
- [x] I have updated relevant documentation to reflect my changes
- [x] I have considered and documented any risks above
- [x] All Paperclip CI gates are green
- [x] Greptile is 5/5 with no open P2s, recommendations, or follow-ups
- [x] I will address all Greptile and reviewer comments before
requesting merge
---------
Co-authored-by: Paperclip <noreply@paperclip.ing>