This PR creates a dedicated function for the agent panel message
editor's placeholder text so that we can wait for the agent
initialization to capture whether they support slash commands or not. On
the one (nice) hand, this allow us to stop matching agents by name and
make this a bit more generic. On the other (bad) hand, the "/ for
commands" bit should take a little second to show up because we can only
know whether an agent supports it after it is initialized.
This is particularly relevant now that we have agents coming from
extensions and for them, we would obviously not be able to match by
name.
Release Notes:
- agent: Fixed agent panel message editor's placeholder text by making
it more accurate as to whether agents support slash commands,
particularly those coming from extensions.
Agents running in extensions need to have a root directory of the
extension's dir for installation and authentication, but *not* for the
conversation itself - otherwise the agent is running things like
terminal commands in the wrong dir.
Release Notes:
- Fixed Agent Server extensions having the current working directory of
the extension rather than the project
This makes it possible to do login via things like `cmd: "node", args:
["my-node-file.js", "login"]`
Also, that command will now use Zed's managed `node` instance.
Release Notes:
- ACP extensions can now run terminal login commands using relative
paths
This PR changes the loading/generating indicator when in the "waiting
for tool call confirmation" state so that's a bit more visible and
discernible as needing your attention, as opposed to a regular
generating state.
<img width="400" alt="Screenshot 2025-11-05 at 10 46@2x"
src="https://github.com/user-attachments/assets/88adbf97-20fb-49c4-9c77-b0a3a22aa14e"
/>
Release Notes:
- agent: Improved the "waiting for confirmation" state visibility so
that you more rapidly know the agent is waiting for you to act.
As discussed with @benbrandt and @mikayla-maki:
* We now tell ACP clients we support the nonstandard `terminal-auth`
`_meta` field for terminal-based authentication
* In the future, we anticipate ACP itself supporting *some* form of
terminal-based authentication, but that hasn't been designed yet or gone
through the RFD process
* For now, this unblocks terminal-based auth
Release Notes:
- Added experimental terminal-based authentication to ACP support
Noticed this whilst testing the Docker debugger. I randomly scrolled the
console off screen and was confused briefly as to why this was the case.
Release Notes:
- The debugger query console will no longer needlessly overscroll.
The button could not be clicked whenever the editor was currently not
focused. This PR fixes this and also registers the action on a more
global level, similar to how this is done for all the other agent
actions.
Release Notes:
- Fixed an issue where the `Expand message editor` button would not work
in agent threads if the message editor was not focused.
This PR makes it clearer that you can click on the file path to open the
corresponding file in the agent panel's "edit bar", which is the element
that shows up in the panel as soon as agent-made edits happen.
Release Notes:
- agent panel: Improved the "go to file" affordance in the edit bar.
Release Notes:
- N/A
---
Previously, agent markdown rendering used hardcoded font sizes
(TextSize::Default and TextSize::Small) which ignored the
agent_ui_font_size and agent_buffer_font_size settings. This updates the
markdown style to respect these settings.
This pull request adds support for customizing the font size of code
blocks in agent responses, making it possible to set a distinct font
size for code within the agent panel. The changes ensure that if the new
setting is not specified, the font size will fall back to the agent UI
font size, maintaining consistent appearance.
(I am a frontend developer without any Rust knowledge so this is
co-authored with Claude Code)
**Theme settings extension:**
* Added a new `agent_buffer_code_font_size` setting to
`ThemeSettingsContent`, `ThemeSettings`, and the default settings JSON,
allowing users to specify the font size for code blocks in agent
responses.
[[1]](diffhunk://#diff-a3bba02a485aba48e8e9a9d85485332378aa4fe29a0c50d11ae801ecfa0a56a4R69-R72)
[[2]](diffhunk://#diff-aed3a9217587d27844c57ac8aff4a749f1fb1fc5d54926ef5065bf85f8fd633aR118-R119)
[[3]](diffhunk://#diff-42e01d7aacb60673842554e30970b4ddbbaee7a2ec2c6f2be1c0b08b0dd89631R82-R83)
* Updated the VSCode import logic to recognize and import the new
`agent_buffer_code_font_size` setting.
**Font size application in agent UI:**
* Modified the agent UI rendering logic in `thread_view.rs` to use the
new `agent_buffer_code_font_size` for code blocks, and to fall back to
the agent UI font size if unset.
[[1]](diffhunk://#diff-f73942e8d4f8c4d4d173d57d7c58bb653c4bb6ae7079533ee501750cdca27d98L5584-R5584)
[[2]](diffhunk://#diff-f73942e8d4f8c4d4d173d57d7c58bb653c4bb6ae7079533ee501750cdca27d98L5596-R5598)
* Implemented a helper method in `ThemeSettings` to retrieve the code
block font size, with fallback logic to ensure a value is always used.
* Updated the settings application logic to propagate the new code block
font size setting throughout the theme system.
### Example Screenshots
---------
Co-authored-by: Danilo Leal <daniloleal09@gmail.com>
Currently Zed only displays agent notifications (e.g. when the agent
completes a task) if the user has switched apps and Zed is not in the
foreground. This adds PR supports the scenario where the agent finishes
a long-running task and the user is busy coding within Zed on something
else.
Releases Note:
- If agent notifications are turned on, they will now also be displayed
when the agent panel is hidden, in complement to them showing when the
Zed window is in the background.
---------
Co-authored-by: Danilo Leal <daniloleal09@gmail.com>
After #38882 we were always including file/directory mentions as
`zed:///agent/file?path=a/b/c.rs`.
However, for most resource links (files/directories/symbols/selections)
we want to use a common format, so that ACP servers don't have to
implement custom handling for parsing `ResourceLink`s coming from Zed.
This is what it looks like now:
```
[@index.js](file:///Users/.../projects/reqwest/examples/wasm_github_fetch/index.js)
[@wasm](file:///Users/.../projects/reqwest/src/wasm)
[@Error](file:///Users/.../projects/reqwest/src/async_impl/client.rs?symbol=Error#L2661:2661)
[@error.rs (23:27)](file:///Users/.../projects/reqwest/src/error.rs#L23:27)
```
Release Notes:
- N/A
---------
Co-authored-by: Cole Miller <cole@zed.dev>
Updates to acp crate 0.7, which allows us to send information about the
client to the Agent.
In the future, we can also use the AgentInfo on the response for
internal metrics.
Release Notes:
- N/A
Previously we had `Context` and `ContextStore` in both `agent_ui` (used
to store context for the inline assistant) and `assistant_context` (used
for text threads) which is confusing.
This PR makes it so that the `assistant_context` concepts are now called
`TextThread*`, the crate was renamed to `assistant_text_thread`
Release Notes:
- N/A
Release Notes:
- agent: Improved the editing previous messages UX by focusing in the
agent panel's message editor after regenerating a prompt, instead of
moving focus to the nearest regular buffer.
Closes#39172
This refactors when we resolve UI keybindings in an effort to reduce
flickering whilst painting these: Previously, we would always resolve
these upon creating the binding. This could lead to cases where the
corresponding context was not yet available and no binding could be
resolved, even if the binding was then available on the next presented
frame. Following that, on the next rerender of whatever requested this
keybinding, the keybind for that context would then be found, we would
render that and then also win a layout shift in that process, as we went
from nothing rendered to something rendered between these frames.
With these changes, this now happens less often, because we only look
for the keybinding once the context can actually be resolved in the
window.
| Before | After |
| --- | --- |
|
https://github.com/user-attachments/assets/adebf8ac-217d-4c7f-ae5a-bab3aa0b0ee8
|
https://github.com/user-attachments/assets/70a82b4b-488f-4a9f-94d7-b6d0a49aada9
|
Also reduced cloning in the keymap editor in this process, since that
requiered changing due to this anyway.
Release Notes:
- Fixed some cases where keybinds would appear with a slight delay,
causing a flicker in the process
Moves the Codex warning into the thread so that we can render it nicer,
as well as provide an option to open the folder in WSL.
Release Notes:
- N/A
---------
Co-authored-by: Bennet Bo Fenner <bennetbo@gmx.de>
We were clearing the message editor too early. We only want to clear the
message editor if we are going to short circuit and return early before
submitting.
Otherwise, the agents that can handle this themselves won't have the
ability to do so.
Release Notes:
- acp: Fix /logout not working for some agents
Previously, if a tool call's output was just text, it would be collapsed
with no way to open it.
Now we track the collapsed cards instead of the expanded ones to allow
all tool calls to be expanded by default, and only collapse the ones
required by settings changes
Release Notes:
- acp: Fix tool call markdown output unintentionally being collapsed by
default
This PR fixes a bug where the review icon button wouldn't properly open
the review tab if you weren't focused in the agent panel's message
editor. The solution was to register the action also at the workspace
level.
Release Notes:
- agent: Fixed a bug where the review icon button wouldn't work to open
the review tab if focus weren't in the panel's message editor.
- Update `AcpThreadView.insert_selections` to take into account whether
the user is currently editing an existing message and, if it is, insert
the selection into that message instead of the thread's message editor
- Update Window's default keymap to use the `agent::QuoteSelection`
action instead of the deprecated `assistant::QuoteSelection` action
- Introduce `AcpThreadView.active_editor` to allow callers to retrieve
either the thread view's message editor or the editor for the message
being edited, in case `AcpThreadView.editing_message` is not `None`
- Improve `AcpThreadView.focus_handle` to focus on the message being
currently edited in case the user navigates back to the editor and then
to the thread view again, all while editing a message
- Add tests for `AcpThreadView.insert_selections`, ensuring that the
selection is inserted in the message being currently edited, if a
message is being edited, or the thread view's message editor if no
message is being edited
Closes#39693
Release Notes:
- Improved `agent: quote selection` to also work for a message that was
already sent but is being edited
---------
Co-authored-by: Ben Brandt <benjamin.j.brandt@gmail.com>
- Prefer agent-specific logout handling to allow state reset
- Treat any auth method as supported; remove provider-specific filter
- Avoid prompting auth when issuing /logout and agent supports it
Release Notes:
- N/A
We don't get an ExitStatus from a remote terminal, so this check was
failing.
Ideally we move all of this to just needing an exit code, but we will
have to revisit that later.
Release Notes:
- N/A
Before this change the active theme and icon theme were retrofitted onto
the ThemeSettings.
Now they're in their own new global (GlobalTheme::theme(cx) and
GlobalTheme::icon_theme(cx))
This lets us remove cx from the settings traits, and tidy up a few other
things along the way.
Release Notes:
- N/A
Fixes a `--release`-only bug in feature-flagged agents where the feature
flag isn't picked up in some situations (unless there was a settings
change to go with it - due to an early return when settings didn't
change).
Release Notes:
- N/A
This only affects `codex-acp` for now.
Not using the PTY in display-only terminals means they don't display the
login prompt (or spurious `%`s) at the end of terminal output
renderings.
Release Notes:
- N/A
Closes https://github.com/zed-industries/zed/issues/39406
Follow up to https://github.com/zed-industries/zed/pull/38726
This PR introduces the `agent_buffer_font_size` setting and renames
`agent_font_size` to `agent_ui_font_size`. This allows whoever wants
`buffer_font_size` and `agent_buffer_font_size` to match, as well as
folks who want a slightly smaller size only in the agent panel (which...
also looks just better by default!).
Release Notes:
- agent: Introduced the `agent_buffer_font_size` setting and renamed
`agent_font_size` to `agent_ui_font_size`, allowing for granular buffer
font size control in the agent panel vs. regular editors.
This was added for Codex, but had undesirable consequences for Claude
Code (on Nightly, never made it to Preview). We're going to address this
in `codex-acp` instead.
Release Notes:
- N/A
Closes#5355
Release Notes:
- Fixed rendering glitches with files with more than 16 million lines
(that occured due to floating number rounding errors).
---------
Co-authored-by: Smit Barmase <heysmitbarmase@gmail.com>
Remove the ad-hoc quoting we were doing before, which only works for
POSIX shells, in favor of using `Shell::WithArguments`.
Release Notes:
- N/A
---------
Co-authored-by: Lukas Wirth <me@lukaswirth.dev>
Closes https://github.com/zed-industries/zed/issues/38690Closes#37353
### Background
On Windows, paths are normally separated by `\`, unlike mac and linux
where they are separated by `/`. When editing code in a project that
uses a different path style than your local system (e.g. remoting from
Windows to Linux, using WSL, and collaboration between windows and unix
users), the correct separator for a path may differ from the "native"
separator.
Previously, to work around this, Zed converted paths' separators in
numerous places. This was applied to both absolute and relative paths,
leading to incorrect conversions in some cases.
### Solution
Many code paths in Zed use paths that are *relative* to either a
worktree root or a git repository. This PR introduces a dedicated type
for these paths called `RelPath`, which stores the path in the same way
regardless of host platform, and offers `Path`-like manipulation APIs.
RelPath supports *displaying* the path using either separator, so that
we can display paths in a style that is determined at runtime based on
the current project.
The representation of absolute paths is left untouched, for now.
Absolute paths are different from relative paths because (except in
contexts where we know that the path refers to the local filesystem)
they should generally be treated as opaque strings. Currently we use a
mix of types for these paths (std::path::Path, String, SanitizedPath).
Release Notes:
- N/A
---------
Co-authored-by: Cole Miller <cole@zed.dev>
Co-authored-by: Piotr Osiewicz <24362066+osiewicz@users.noreply.github.com>
Co-authored-by: Peter Tripp <petertripp@gmail.com>
Co-authored-by: Smit Barmase <heysmitbarmase@gmail.com>
Co-authored-by: Lukas Wirth <me@lukaswirth.dev>
Closes https://github.com/zed-industries/zed/issues/38528
In the agent panel's `thread_view.rs` file, we have a `render_tool_call`
function that controls what we show in the UI for most types of tools.
However, for some of them—for example, terminal/execute and edit
tools—we have a special rendering so we can tailor the UI for their
specific needs. But... before the specific rendering function is called,
all tools still go through the `render_tool_call`.
Problem is that, in the case of the terminal tool, you couldn't see the
full command the agent wants to run when the tool is still in its
`render_tool_call` state. That's mostly because of the treatment we give
to labels while in that state. A particularly bad scenario because
well... seeing the _full_ command _before_ you choose to accept or
reject is rather important.
This PR fixes that by essentially special-casing the terminal tool
display when in the `render_tool_call` rendering state, so to speak.
There's still a slight UI misalignment I want to fix but it shouldn't
block this fix to go out.
Here's our final result:
<img width="400" height="1172" alt="Screenshot 2025-09-23 at 6 19@2x"
src="https://github.com/user-attachments/assets/71c79e45-ab66-4102-b046-950f137fa3ea"
/>
Release Notes:
- agent: Fixed terminal command not being fully displayed while in the
"waiting for confirmation" state.
Nowadays, people don't expect @-mentioning a directory to include the
contents of all files within it. Doing so makes it very likely to
consume an undesirable amount of tokens.
By default, we'll now only include the path of the directory and let the
model decide how much to read via tools. We'll still include the
contents if no tools are available (e.g. "Minimal" profile is selected).
Release Notes:
- Agent Panel: Do not include the content of @-mentioned directories
when tools are available
Danilo Leal
DavidandDanilo Leal
Mikayla MakiandNia
Bennet Bo Fenner
Richard Feldman
Lukas Wirth
Finn Evers
Mark ChristiansenandDanilo Leal
Charles McLaughlinandDanilo Leal
Ben Brandt
DinoandBen Brandt
Conrad Irwin
Piotr OsiewiczandSmit Barmase
Agus Zubiaga