Commit Graph
100 Commits
Author SHA1 Message Date
Agus Zubiaga 18ac4ac5ef ACP debug tools pane (#36768)
Adds a new "acp: open debug tools" action that opens a new workspace
item with a log of ACP messages for the active connection.

Release Notes:

- N/A
2025-08-22 19:32:49 +00:00
Agus Zubiaga 0beb919bbb acp: Fix MessageEditor::set_message for sent messages (#36715)
The `PromptCapabilities` introduced in previous PRs were only getting
set on the main message editor and not for the editors in user messages.
This caused a bug where mentions would disappear after resending the
message, and for the completion provider to be limited to files.

Release Notes:

- N/A
2025-08-21 17:29:53 -03:00
Agus ZubiagaandBennet Bo Fenner 2234f91b7b acp: Remove invalid creases on edit (#36708)
Release Notes:

- N/A

Co-authored-by: Bennet Bo Fenner <bennetbo@gmx.de>
2025-08-21 16:56:40 -03:00
Agus ZubiagaandBennet Bo Fenner 725ed5dd01 acp: Hide loading diff animation for external agents and update in place (#36699)
The loading diff animation can be jarring for external agents because
they stream the diff at the same time the tool call is pushed, so it's
only displayed while we're asynchronously calculating the diff. We'll
now only show it for the native agent.

Also, we'll now only update the diff when it changes, which avoids
unnecessarily hiding it for a few frames.

Release Notes:

- N/A

Co-authored-by: Bennet Bo Fenner <bennetbo@gmx.de>
2025-08-21 16:56:15 -03:00
Agus Zubiaga 4bee06e507 acp: Use ResourceLink for agents that don't support embedded context (#36687)
The completion provider was already limiting the mention kinds according
to `acp::PromptCapabilities`. However, it was still using
`ContentBlock::EmbeddedResource` when
`acp::PromptCapabilities::embedded_context` was `false`. We will now use
`ResourceLink` in that case making it more complaint with the
specification.

Release Notes:

- N/A
2025-08-21 14:57:46 +00:00
Agus Zubiaga 9a3e4c47d0 acp: Suggest upgrading to preview instead of latest (#36648)
A previous PR changed the install command from `@latest` to `@preview`,
but the upgrade command kept suggesting `@latest`.

Release Notes:

- N/A
2025-08-21 00:52:38 +00:00
Agus Zubiaga 6f242772cc acp: Update to 0.0.30 (#36643)
See: https://github.com/zed-industries/agent-client-protocol/pull/20

Release Notes:

- N/A
2025-08-21 00:10:36 +00:00
Agus Zubiaga 74c0ba980b acp: Reliably suppress gemini abort error (#36640)
https://github.com/zed-industries/zed/pull/36633 relied on the prompt
request responding before cancel, but that's not guaranteed


Release Notes:

- N/A
2025-08-20 23:32:17 +00:00
Agus Zubiaga ffb995181e acp: Supress gemini aborted errors (#36633)
This PR adds a temporary workaround to supress "Aborted" errors from
Gemini when cancelling generation. This won't be needed once
https://github.com/google-gemini/gemini-cli/pull/6656 is generally
available.

Release Notes:

- N/A
2025-08-20 22:30:25 +00:00
Agus Zubiaga 9e34bb3f05 acp: Hide feedback buttons for external agents (#36630)
Release Notes:

- N/A
2025-08-20 21:35:48 +00:00
Agus Zubiaga d1820b183a acp: Suggest installing gemini@preview instead of latest (#36629)
Release Notes:

- N/A
2025-08-20 21:26:07 +00:00
Agus Zubiaga 02dabbb9fa acp thread view: Do not go into editing mode if unsupported (#36623)
Release Notes:

- N/A
2025-08-20 20:05:53 +00:00
Agus Zubiaga 2813073d7b message editor: Only allow types of content the agent can handle (#36616)
Uses the new
[`acp::PromptCapabilities`](https://github.com/zed-industries/agent-client-protocol/blob/a39b7f635d67528f0a4e05e086ab283b9fc5cb93/rust/agent.rs#L194-L215)
to disable non-file mentions and images for agents that don't support
them.

Release Notes:

- N/A
2025-08-20 19:04:10 +00:00
Agus Zubiaga b12d862236 Rename acp flag (#36541)
Release Notes:

- N/A
2025-08-20 02:11:17 +00:00
Agus Zubiaga 7c7043947b Improve claude tools (#36538)
- Return unified diff from `Edit` tool so model can see the final state
- Format on save if enabled
- Provide `Write` tool
- Disable `MultiEdit` tool
- Better prompting

Release Notes:

- N/A
2025-08-20 01:42:11 +00:00
Agus Zubiaga 714c36fa7b claude: Include all mentions and images in user message (#36539)
User messages sent to Claude Code will now include the content of all
mentions, and any images included.

Release Notes:

- N/A
2025-08-19 22:30:26 -03:00
Agus Zubiaga e092aed253 Split external agent flags (#36499)
Release Notes:

- N/A
2025-08-19 14:25:25 -03:00
Agus Zubiaga b578031120 claude: Respect always allow setting (#36450)
Claude will now respect the `agent.always_allow_tool_actions` setting
and will set it when "Always Allow" is clicked.

Release Notes:

- N/A
2025-08-18 23:27:08 +00:00
Agus Zubiaga 8b89ea1a80 Handle auth for claude (#36442)
We'll now use the anthropic provider to get credentials for `claude` and
embed its configuration view in the panel when they are not present.

Release Notes:

- N/A
2025-08-18 20:40:59 +00:00
Agus ZubiagaandDanilo 48fed866e6 acp: Have AcpThread handle all interrupting (#36417)
The view was cancelling the generation, but `AcpThread` already handles
that, so we removed that extra code and fixed a bug where an update from
the first user message would appear after the second one.

Release Notes:

- N/A

Co-authored-by: Danilo <danilo@zed.dev>
2025-08-18 12:34:27 -03:00
Agus Zubiaga db31fa67f3 acp: Stay in edit mode when current completion ends (#36413)
When a turn ends and the checkpoint is updated, `AcpThread` emits
`EntryUpdated` with the index of the user message. This was causing the
message editor to be recreated and, therefore, lose focus.

Release Notes:

- N/A
2025-08-18 11:37:28 -03:00
Agus ZubiagaandDanilo Leal e2db434920 acp thread view: Floating editing message controls (#36283)
Prevents layout shift when focusing the editor

Release Notes:

- N/A

---------

Co-authored-by: Danilo Leal <daniloleal09@gmail.com>
2025-08-18 12:50:29 +00:00
Agus Zubiaga 9eb1ff2726 acp thread view: Always use editors for user messages (#36256)
This means the cursor will be at the position you clicked:


https://github.com/user-attachments/assets/0693950d-7513-4d90-88e2-55817df7213a


Release Notes:

- N/A
2025-08-15 21:03:36 +00:00
Agus Zubiaga 91e6b38285 Log agent servers stderr (#36243)
Release Notes:

- N/A
2025-08-15 10:58:57 -03:00
Agus Zubiaga 43ee604179 acp: Clean up entry views on rewind (#36197)
We were leaking diffs and terminals on rewind, we'll now clean them up.
This PR also introduces a refactor of how we mantain the entry view
state to use a `Vec` that's kept in sync with the thread entries.

Release Notes:

- N/A
2025-08-14 18:30:18 +00:00
Agus ZubiagaandCole Miller 389d24d7e5 Fully support all mention kinds (#36134)
Feature parity with the agent1 @mention kinds:
- File
- Symbols
- Selections
- Threads
- Rules
- Fetch


Release Notes:

- N/A

---------

Co-authored-by: Cole Miller <cole@zed.dev>
2025-08-13 20:11:32 +00:00
Agus ZubiagaandCole Miller 44953375cc Include mention context in acp-based native agent (#36006)
Also adds data-layer support for symbols, thread, and rules.

Release Notes:

- N/A

---------

Co-authored-by: Cole Miller <cole@zed.dev>
2025-08-12 13:12:58 +00:00
2526dcb5a5 agent2: Port edit_file tool (#35844)
TODO:
- [x] Authorization
- [x] Restore tests

Release Notes:

- N/A

---------

Co-authored-by: Antonio Scandurra <me@as-cii.com>
Co-authored-by: Ben Brandt <benjamin.j.brandt@gmail.com>
2025-08-08 12:43:53 +00:00
Agus Zubiaga edef1f1470 Fix acp generating status after stop (#35852)
Release Notes:

- N/A
2025-08-08 05:26:53 +00:00
Agus Zubiaga 3d662ee282 agent2: Port read_file tool (#35840)
Ports the read_file tool from `assistant_tools` to `agent2`. 

Note: Image support not implemented.

Release Notes:

- N/A
2025-08-07 20:46:47 -03:00
Agus Zubiaga 6912dc8399 Fix CC tool state on cancel (#35763)
When we stop the generation, CC tells us the tool completed, but it was
actually cancelled.

Release Notes:

- N/A
2025-08-07 20:26:19 -03:00
Agus ZubiagaandCole Miller bd1c26cb5b Fix interrupting ACP threads and CC cancellation (#35752)
Fixes a bug where generation wouldn't continue after interrupting the
agent, and improves CC cancellation so we don't display "[Request
interrupted by user]"

Release Notes:

- N/A

---------

Co-authored-by: Cole Miller <cole@zed.dev>
2025-08-06 22:55:17 -03:00
Agus Zubiaga 250c51bb20 Fix syntax highlighting in ACP diffs (#35748)
Release Notes:

- N/A
2025-08-06 19:53:45 +00:00
Agus ZubiagaandDavid Kleingeld 33f198fef1 Thread view scrollbar (#35655)
This also adds a convenient `Scrollbar:auto_hide` function so that we
don't have to handle that at the callsite.

Release Notes:

- N/A

---------

Co-authored-by: David Kleingeld <davidsk@zed.dev>
2025-08-06 14:01:34 +00:00
Agus Zubiaga 334bdd0efc Fix acp thread entry width (#35723)
Release Notes:

- N/A
2025-08-06 13:39:55 +00:00
Agus Zubiaga 69dc870828 Fix CC todo tool parsing (#35721)
It looks like the TODO tool call no longer requires a priority.

Release Notes:

- N/A
2025-08-06 13:27:11 +00:00
Agus Zubiaga 22fa41e9c0 Handle CC thinking (#35722)
Release Notes:

- N/A
2025-08-06 13:20:53 +00:00
Agus Zubiaga 3bbd32b70e Support CC migrate-installer path (#35717)
If we can't find CC in the PATH, we'll now fall back to a known local
install path.

Release Notes:

- N/A
2025-08-06 12:23:47 +00:00
Agus Zubiaga b7469f5bc3 Fix ACP connection and thread leak (#35670)
When you switched away from an ACP thread, the `AcpThreadView` entity
(and thus thread, and subprocess) was leaked. This happened because we
were using `cx.processor` for the `list` state callback, which uses a
strong reference.

This PR changes the callback so that it holds a weak reference, and adds
some tests and assertions at various levels to make sure we don't
reintroduce the leak in the future.

Release Notes:

- N/A
2025-08-05 19:10:51 -03:00
Agus Zubiaga 1325bf1420 Update to acp 0.0.18 (#35595)
Release Notes:

- N/A
2025-08-04 18:45:17 +00:00
Agus Zubiaga a6a34dad0f Fix gemini e2e tests (#35583)
Release Notes:

- N/A
2025-08-04 15:02:22 +00:00
Agus Zubiaga 9353ba7887 Fix remaining agent server integration tests (#35222)
Release Notes:

- N/A
2025-07-29 12:40:59 +00:00
Agus Zubiaga fd68265efd Fix integration tests for claude (#35212)
Release Notes:

- N/A
2025-07-28 16:18:01 +00:00
Agus ZubiagaandBen Brandt c2fc70eef7 ACP over MCP server impl (#35196)
Release Notes:

- N/A

---------

Co-authored-by: Ben Brandt <benjamin.j.brandt@gmail.com>
2025-07-28 14:14:10 +00:00
Agus Zubiaga af0c909924 McpServerTool output schema (#35069)
Add an `Output` associated type to `McpServerTool`, so that we can
include its schema in `tools/list`.

Release Notes:

- N/A
2025-07-25 02:57:18 +00:00
Agus Zubiaga 15c9da4ea4 Add ability to register tools in McpServer (#35068)
Makes it easier to add tools to a server by implementing a trait

Release Notes:

- N/A
2025-07-25 02:19:20 +00:00
2d0f10c48a Refactor to use new ACP crate (#35043)
This will prepare us for running the protocol over MCP

Release Notes:

- N/A

---------

Co-authored-by: Ben Brandt <benjamin.j.brandt@gmail.com>
Co-authored-by: Conrad Irwin <conrad.irwin@gmail.com>
Co-authored-by: Richard Feldman <oss@rtfeldman.com>
2025-07-24 14:39:29 -03:00
Agus Zubiaga 5d985fa1d8 Improve MCP server responses (#34927)
Release Notes:

- N/A
2025-07-22 22:14:34 +00:00
Agus ZubiagaandBen Brandt 722a05bc21 Wire up stop button in claude threads (#34839)
Release Notes:

- N/A

---------

Co-authored-by: Ben Brandt <benjamin.j.brandt@gmail.com>
2025-07-21 20:33:59 +00:00
Agus Zubiaga cc56196152 Fix loading agent server settings (#34662)
Release Notes:

- N/A
2025-07-21 14:26:00 +00:00
Agus ZubiagaandDanilo Leal 405244d422 Display ACP plans (#34816)
Release Notes:

- N/A

---------

Co-authored-by: Danilo Leal <daniloleal09@gmail.com>
2025-07-21 11:11:37 -03:00
Agus Zubiaga dab0b3509d Unify agent server settings and extract e2e tests out (#34642)
Release Notes:

- N/A
2025-07-17 17:34:38 +00:00
8e4555455c Claude experiment (#34577)
Release Notes:

- N/A

---------

Co-authored-by: Conrad Irwin <conrad.irwin@gmail.com>
Co-authored-by: Anthony Eid <hello@anthonyeid.me>
Co-authored-by: Ben Brandt <benjamin.j.brandt@gmail.com>
Co-authored-by: Nathan Sobo <nathan@zed.dev>
Co-authored-by: Oleksiy Syvokon <oleksiy.syvokon@gmail.com>
2025-07-17 14:25:55 +00:00
Agus Zubiaga 6f6c2915b2 Display and jump to tool locations (#34304)
Release Notes:

- N/A
2025-07-11 14:52:21 -03:00
Agus ZubiagaandBen Brandt d0e01dbd8f Improve thread message history (#34299)
- Keep history across threads
- Reset position when edited

Release Notes:

- N/A

---------

Co-authored-by: Ben Brandt <benjamin.j.brandt@gmail.com>
2025-07-11 16:24:41 +00:00
Agus ZubiagaandConrad Irwin 08ffd9884a Update to acp 0.0.6 (#34159)
Release Notes:

- N/A

---------

Co-authored-by: Conrad Irwin <conrad.irwin@gmail.com>
2025-07-09 19:14:29 -06:00
8a1e795746 agent: Move ActiveThread and MessageEditor into ActiveView (#33462)
`ActiveThread` and `MessageEditor` only make sense when `active_view` is
`Thread`, so we moved them in there. This will make it easier to work on
new agent threads.

Release Notes:

- N/A

---------

Co-authored-by: Antonio Scandurra <me@as-cii.com>
Co-authored-by: Conrad Irwin <conrad.irwin@gmail.com>
Co-authored-by: Smit Barmase <heysmitbarmase@gmail.com>
2025-06-26 15:36:23 +00:00
Agus Zubiaga b103d7621b Improve handling of large output in embedded terminals (#32416)
#31922 made embedded terminals automatically grow to fit the content. We
since found some issues with large output which this PR addresses by:

- Only shaping / laying out lines that are visible in the viewport
(based on `window.content_mask`)
- Falling back to embedded scrolling after 1K lines. The perf fix above
actually makes it possible to handle a lot of lines, but:
- Alacrity uses a `u16` for rows internally, so we needed a limit to
prevent overflow.
- Scrolling through thousands of lines to get to the other side of a
terminal tool call isn't great UX, so we might as well set the limit
low.
- We can consider raising the limit when we make card headers sticky.

Release Notes:

- Agent: Improve handling of large terminal output
2025-06-09 18:11:31 -03:00
Agus Zubiaga b7abc9d493 agent: Display full terminal output without scrolling (#31922)
The terminal tool card used a fixed height and scrolling, but this meant
that it was too tall for commands that only outputted a few lines, and
the nested scrolling was undesirable.

This PR makes the card be as too as needed to fit the entire output (no
scrolling), and allows the user to collapse it to fewer lines when
applicable. Making it work the same way as the edit tool card. In fact,
both tools now use a shared UI component.


https://github.com/user-attachments/assets/1127e21d-1d41-4a4b-a99f-7cd70fccbb56


Release Notes:

- Agent: Display full terminal output
- Agent: Allow collapsing terminal output
2025-06-03 10:54:25 -07:00
Agus Zubiaga bb9e2b0403 Fix existing CompletionMode deserialization (#31790)
https://github.com/zed-industries/zed/pull/31668 renamed
`CompletionMode::Max` to `CompletionMode::Burn` which is a good change,
but this broke the deserialization for threads whose completion mode was
stored in LMDB. This adds a deserialization alias so that both values
work.

We could make a full new `SerializedThread` version which migrates this
value, but that seems overkill for this single change, we can batch that
with more changes later. Also, people in nightly already have some v1
threads with `burn` stored, so it wouldn't quite work for everybody.

Release Notes:

- N/A
2025-05-30 19:06:08 -03:00
Agus Zubiaga b1c7fa1dac Debounce language server file system events (#30773)
This helps prevent a race condition where the language server would
update in the middle of a `git checkout`

Release Notes:

- N/A
2025-05-20 09:00:28 +00:00
Agus Zubiaga a4766e296f Add tool result image support to Gemini models (#30647)
Release Notes:

- Add tool result image support to Gemini models
2025-05-14 00:51:31 +00:00
Agus Zubiaga dd6594621f Add image input support for OpenAI models (#30639)
Release Notes:

- Added input image support for OpenAI models
2025-05-13 17:32:42 +02:00
Agus ZubiagaandRichard Feldman f0da3b74f8 agent: Handle thread title generation errors (#30273)
The title of a (text) thread would get stuck in "Loading Summary..."
when the request to generate it failed. We now handle this case by
falling back to the default title, and letting the user manually edit
the title or retry generating it.



https://github.com/user-attachments/assets/898d26ad-d31f-4b62-9b05-519d923b1b22



Release Notes:

- agent: Handle thread title generation errors

---------

Co-authored-by: Richard Feldman <oss@rtfeldman.com>
2025-05-12 07:45:48 +00:00
82d271cb5b agent: Thread history update improvements (#30415)
- Try to preserve previously selected item on update
- Do not clear list items while updating to avoid a frame with no items
rendered

Release Notes:

- agent: Preserve previously selected item in Thread History on update

---------

Co-authored-by: Danilo Leal <daniloleal09@gmail.com>
Co-authored-by: Richard Feldman <oss@rtfeldman.com>
2025-05-09 16:13:50 -03:00
Agus Zubiaga c512d43e8c agent: Render edit tool error as markdown (#30325)
Release Notes:

- agent: Render edit tool error as markdown and allow selecting it
2025-05-09 01:18:52 +00:00
Agus Zubiaga 9810745465 agent: Fix autoscrolling to history entry (#30321)
We were still using entry indexes to scroll, but the list now includes
the separators as items, so the indexes need to be translated

Release Notes:

- agent: Fix autoscrolling to history entry when navigating via keyboard
2025-05-08 20:48:57 -03:00
Agus Zubiaga 4b61d4ba6a agent: Fix message editor's button positions when expanded (#30311)
Fixes an issue introduced by #29959 which caused the message editor to
overflow from the agent panel bounds, making the bottom buttons
invisible when the editor was expanded (cmd+esc).

Fixing this required changing the base structure of the agent panel, but
things seem to work as expected:


https://github.com/user-attachments/assets/fc4c97fb-f7cb-4f54-a268-c30fbcb1649f


Release Notes:

- agent: Fix message editor's button positions when expanded
2025-05-08 20:36:19 -03:00
Agus Zubiaga 77ac82587a agent: Improve consecutive tool use callout spacing (#30145)
Release Notes:

- agent: Fix "consecutive tool use limit" callout spacing
2025-05-07 17:43:04 +00:00
Agus Zubiaga 264097e253 agent: Use correct timezone for thread history separators (#30059)
Turns out `naive_local` doesn't actually offset a `DateTime<Utc>` to the
local timezone before creating a `NaiveDate`.

Release Notes:

- agent: Use correct timezone for thread history separators
2025-05-07 04:41:04 +00:00
3cdf5ce947 agent: Allow customizing temperature by provider/model (#30033)
Adds a new `agent.model_parameters` setting that allows the user to
specify a custom temperature for a provider AND/OR model:

```json5
    "model_parameters": [
      // To set parameters for all requests to OpenAI models:
      {
        "provider": "openai",
        "temperature": 0.5
      },
      // To set parameters for all requests in general:
      {
        "temperature": 0
      },
      // To set parameters for a specific provider and model:
      {
        "provider": "zed.dev",
        "model": "claude-3-7-sonnet-latest",
        "temperature": 1.0
      }
    ],
```

Release Notes:

- agent: Allow customizing temperature by provider/model

---------

Co-authored-by: Max Brunsfeld <maxbrunsfeld@gmail.com>
Co-authored-by: Marshall Bowers <git@maxdeviant.com>
2025-05-06 20:36:25 +00:00
Agus Zubiaga de554589a8 agent: Add date separators to Thread History (#29961)
Adds time-bucket separators to the thread history list:


https://github.com/user-attachments/assets/c9ac3ec4-b632-4ea5-8234-382b48de2bd6

Note: I'm simulating that Today is next Thursday so that I can show the
"This Week" bucket.

Release Notes:

- agent: Add date separators to Thread History
2025-05-06 10:18:48 +00:00
Agus Zubiaga 9c11d24887 Fix hiding editor toolbar and add agent_review setting (#29854)
Closes #29836

The agent diff toolbar item was causing the editor toolbar to show even
when all the other elements were disabled via settings.

This PR fixes this by setting the location to
`ToolbarItemLocation::Hidden` in the states where it shouldn't show.

It also adds a new a `toolbar.agent_review` setting to hide the agent
review buttons altogether. However, if the other toolbar elements are
hidden and the file isn't under review, the editor toolbar will still be
hidden. So you only need to set this to `false` if you don't want them
to show up even under agent review.

Release Notes:

- N/A
2025-05-03 17:43:46 -03:00
Agus Zubiaga 1877fce609 agent: Fix default cursor position on reviewing editors (#29825)
The cursor wasn't always placed at the first hunk for review editors.

Release Notes:

- N/A
2025-05-02 21:58:00 +00:00
Agus Zubiaga 64316309aa agent: Review edits in single-file editors (#29820)
Enables reviewing agent edits from single-file editors in addition to
the multibuffer experience we already had.


https://github.com/user-attachments/assets/a2c287f0-51d6-43a1-8537-821498b91983


This feature can be turned off by setting `assistant.single_file_review:
false`.

Release Notes:

- agent: Review edits in single-file editors
2025-05-02 17:57:16 -03:00
Agus Zubiaga 2508e491d5 agent: Discourage long-running commands (#29627)
Adds popular examples of long-running commands to system prompt. 

Unfortunately, I couldn't add an eval example as the new terminal tool
no longer works in `eval`. We can look into that tomorrow, but I'm
seeing improvements when manually testing this, so I'd like to merge it.

<img
src="https://github.com/user-attachments/assets/ac24e617-e068-466f-875d-c30e1f2465c4"
width=400></img>


Release Notes:

- agent: Discourage long-running commands
2025-04-29 19:21:16 -03:00
Agus Zubiaga fd17f2d8ae agent: Enrich grep tool output with syntax information (#29601)
The `grep` tool used to include 4 lines of context around the match, but
the lines included would often be unhelpful. This PR improves this
behavior by using the range of the parent syntax node that contains the
full line(s) matched.

The match headers will also now include symbol breadcrumbs so that the
model can already gather code structure before/without reading files.

````md
### impl GitRepository for RealGitRepository › fn compare_checkpoints › L1278-1284
```rust
                let result = git
                    .run(&[
                        "diff-tree",
                        "--quiet",
                        &left.commit_sha.to_string(),
                        &right.commit_sha.to_string(),
                    ])
```
````

This positively impacts the `add_arg_to_trait_method` eval example with
better diff output, fewer tool failures, and reduced total turns.

Note: We have some plans to use a an "elision" approach where we would
combine all matches for a given file, skipping lines between them while
keeping symbol declaration lines. The theory is that this would be map
more closely to the expected input for edits. For now, this PR is a
significant improvement.

Release Notes:

- Agent: Enrich `grep` tool output with syntax information
2025-04-29 17:03:02 +00:00
Agus Zubiaga c157b1c455 rules: How to run clippy (#29247)
Tell model how to run clippy in `.rules`

Release Notes:

- N/A
2025-04-25 13:25:39 -03:00
Agus ZubiagaandBennet Bo Fenner 58604fba86 agent: Do not reuse assistant message across generations (#29360)
#29354 introduced a bug where we would append tool uses to the last
assistant message even if it was from a previous request.

Release Notes:

- N/A

Co-authored-by: Bennet Bo Fenner <bennetbo@gmx.de>
2025-04-24 17:56:47 +00:00
f81e65ae7c agent: Do not create user messages for tool results in thread (#29354)
We used to insert empty user messages into the `Thread::messages` `Vec`
when tools finished running and then we would attach the results when
creating the request. This approach was very easy to mess up during
state handling, leading to empty user messages displayed in the
conversation and API failures.

Instead, we will no longer insert actual user messages for tool results
to the `Thread`, and will only do this on the fly when creating the
model request. This simplifies a lot of code and show fix the mentioned
errors.

Release Notes:

- agent: Improve reliability of LLM requests when including tool results

---------

Co-authored-by: Bennet Bo Fenner <bennetbo@gmx.de>
Co-authored-by: Oleksiy Syvokon <oleksiy.syvokon@gmail.com>
2025-04-24 16:30:15 +00:00
Agus Zubiaga 8b5835de17 agent: Improve initial file search quality (#29317)
This PR significantly improves the quality of the initial file search
that occurs when the model doesn't yet know the full path to a file it
needs to read/edit.

Previously, the assertions in file_search often failed on main as the
model attempted to guess full file paths. On this branch, it reliably
calls `find_path` (previously `path_search`) before reading files.

After getting the model to find paths first, I noticed it would try
using `grep` instead of `path_search`. This motivated renaming
`path_search` to `find_path` (continuing the analogy to unix commands)
and adding system prompt instructions about proper tool selection.

Note: I know the command is just called `find`, but that seemed too
general.

In my eval runs, the `file_search` example improved from 40% ± 10% to
98% ± 2%. The only assertion I'm seeing occasionally fail is "glob
starts with `**` or project". We can probably add some instructions in
that regard.

Release Notes:

- N/A
2025-04-23 21:24:41 -03:00
Agus Zubiaga 2124b7ea99 agent: Encourage model to include displayed fields first (#29308)
Instructs the model to include the fields that we display first in the
input object, so that e.g the user can see the path of a file while the
model generates the content.

Release Notes:

- N/A
2025-04-23 20:16:15 -03:00
Agus ZubiagaandRichard Feldman 45d3f5168a eval: New add_arg_to_trait_method example (#29297)
Release Notes:

- N/A

---------

Co-authored-by: Richard Feldman <oss@rtfeldman.com>
2025-04-23 18:46:39 +00:00
ce1a674eba eval: Fine-grained assertions (#29246)
- Support programmatic examples
([example](https://github.com/zed-industries/zed/blob/17feb260a0919e102ae4e220669c467885bf4b57/crates/eval/src/examples/file_search.rs))
- Combine data-driven example declarations into a single `.toml` file
([example](https://github.com/zed-industries/zed/blob/17feb260a0919e102ae4e220669c467885bf4b57/crates/eval/src/examples/find_and_replace_diff_card.toml))
- Run judge on individual assertions (previously called "criteria")
- Report judge and programmatic assertions in one combined table

Note: We still need to work on concept naming 

<img width=400
src="https://github.com/user-attachments/assets/fc719c93-467f-412b-8d47-68821bd8a5f5">

Release Notes:

- N/A

---------

Co-authored-by: Richard Feldman <oss@rtfeldman.com>
Co-authored-by: Max Brunsfeld <maxbrunsfeld@gmail.com>
Co-authored-by: Thomas Mickley-Doyle <tmickleydoyle@gmail.com>
2025-04-22 23:58:58 -03:00
Agus Zubiaga b14356d1d3 agent: Do not add <using_tool> placeholder (#29194)
Our provider code in `language_models` filters out messages for which
`LanguageModelRequestMessage::contents_empty` returns `false`. This
doesn't seem wrong by itself, but `contents_empty` was returning `false`
for messages whose first segment didn't contain non-whitespace text even
if they contained other non-empty segments. This caused requests to fail
when a message with a tool call didn't contain any preceding text.

Release Notes:

- N/A
2025-04-22 00:41:47 -03:00
Agus Zubiaga be76942a69 agent: Migrate tool names in settings (#29168)
Release Notes:

- agent: Add migration to rename `find_replace_file` tool to
`edit_file`, and `regex_search` to `grep`.
2025-04-21 17:03:42 +00:00
Agus Zubiaga 3a27e8c311 edit tool: Handle over-indentation in replace_with_flexible_indent (#29153)
Release Notes:

- agent: Correct over-indentation in search/replace strings from model
2025-04-21 11:02:08 -03:00
Agus Zubiaga 4473b45c3d inline assistant: Fix model picker (#29136)
Release Notes:

- inline assistant: Fixed a bug where the default model would be used
even when a specific inline assistant model was configured
2025-04-21 01:12:57 +00:00
Agus Zubiaga ceeae790b7 eval: Improve lang server idle detection (#29135)
Brings back #29013 after it was accidentally reverted by
https://github.com/zed-industries/zed/pull/28961/commits/e9bb15b9063615762c866c30aaf646acb12af1f3.

Release Notes:

- N/A
2025-04-21 00:17:28 +00:00
Agus Zubiaga f07695c4cd Remove evals crate (#28968)
Release Notes: 
- N/A
2025-04-17 14:53:22 +00:00
Agus Zubiaga a7a7335da4 edit prediction: Assign providers when client status changes (#28919)
There was recently a change that caused the Zed Edit Prediction provider
to only be assigned when the client was connected. However, this check
happened too early, resulting in restored buffers never getting
registered. We'll now subscribe to client status changes and reassign
providers accordingly.

Release Notes:

- edit prediction: Fixed bug disabling prediction in restored buffers
2025-04-16 21:09:25 -03:00
0286b8ab3e agent: Fix conversation token usage and estimate unsent message (#28878)
The UI was mistakenly using the cumulative token usage for the token
counter. It will now display the last request token count, plus an
estimation of the tokens in the message editor and context entries that
haven't been sent yet.


https://github.com/user-attachments/assets/0438c501-b850-4397-9135-57214ca3c07a

Additionally, when the user edits a message, we'll display the actual
token count up to it and estimate the tokens in the new message.

Note: We don't currently estimate the delta when switching profiles. In
the future, we want to use the count tokens API to measure every part of
the request and display a breakdown.

Release Notes:

- agent: Made the token count more accurate and added back estimation of
used tokens as you type and add context.

---------

Co-authored-by: Bennet Bo Fenner <bennetbo@gmx.de>
Co-authored-by: Danilo Leal <daniloleal09@gmail.com>
2025-04-16 16:27:36 -03:00
Agus Zubiaga 0182e09e33 eval: Do not create run files for skipped examples (#28800)
Release Notes:

- N/A
2025-04-15 18:00:04 +00:00
Agus ZubiagaandOleksiy ff4334efc7 eval: Fix stalling on tool confirmation (#28786)
The `always_allow_tool_actions` setting would get overridden with the
default when we loaded each example project, leading to examples
stalling when they run a tool that needed confirmation. There's now a
separate `runner_settings.json` file where we can configure the
environment for the eval.

Release Notes:

- N/A

---------

Co-authored-by: Oleksiy <oleksiy@zed.dev>
2025-04-15 16:53:45 +00:00
Agus Zubiaga e4cf7fe8f5 eval: Improve readability with colors and alignment (#28761)
![CleanShot 2025-04-15 at 10 35
39@2x](https://github.com/user-attachments/assets/495d96fb-fe2f-478b-a9d6-678c1184db9a)


Release Notes:

- N/A
2025-04-15 13:50:01 +00:00
Agus Zubiaga b45230784d agent: Handle context window exceeded errors from Anthropic (#28688)
![CleanShot 2025-04-14 at 11 15
38@2x](https://github.com/user-attachments/assets/9e803ffb-74fd-486b-bebc-2155a407a9fa)

Release Notes:

- agent: Handle context window exceeded errors from Anthropic
2025-04-14 14:39:33 +00:00
Agus ZubiagaandMichael dafe994eef agent: Register tracked buffers with language servers (#28610)
Release Notes:

- agent: Start language servers when accessing files via tools

Co-authored-by: Michael <michael@zed.dev>
2025-04-11 22:27:24 +00:00
90bcde116f agent: Use current shell (#28470)
Release Notes:

- agent: Replace `bash` tool with `terminal` tool which uses the current
shell

---------

Co-authored-by: Bennet <bennet@zed.dev>
Co-authored-by: Antonio <antonio@zed.dev>
2025-04-09 23:38:36 -06:00
Agus Zubiaga 1cb4f8288d Fix bash tool output (#28391) 2025-04-09 08:20:24 -06:00
Agus ZubiagaandMichael Sloan 85c5d8af3a agent: Truncate bash tool output (#28291)
The bash tool will now truncate its output to 8192 bytes (or the last
newline before that).

We also added a global limit for any tool that produces a clearly large
output that wouldn't fit the context window.

Release Notes:

- agent: Truncate bash tool output

---------

Co-authored-by: Michael Sloan <mgsloan@gmail.com>
2025-04-08 16:55:35 +00:00