Conversation
Summary of ChangesHello @gundermanc, I'm Gemini Code Assist1! I'm currently reviewing this pull request and will post my feedback shortly. In the meantime, here's a summary to help you and other reviewers quickly get up to speed! This pull request significantly upgrades the user experience for interacting with subagents by implementing a structured, real-time progress display. Instead of basic text output, the CLI now renders a dynamic view of subagent operations, including their thoughts, tool executions, and any encountered errors. This change provides users with clearer insights into the subagent's current state and actions, making the interaction more transparent and informative. Highlights
Using Gemini Code AssistThe full guide for Gemini Code Assist can be found on our documentation page, here are some quick tips. Invoking Gemini You can request assistance from Gemini at any point by creating a comment using either
Customization To customize Gemini Code Assist for GitHub experience, repository maintainers can create a configuration file and/or provide a custom code review style guide (such as PEP-8 for Python) by creating and adding files to a Limitations & Feedback Gemini Code Assist may make mistakes. Please leave feedback on any instances where its feedback is incorrect or counter productive. You can react with 👍 and 👎 on @gemini-code-assist comments. If you're interested in giving your feedback about your experience with Gemini Code Assist for Github and other Google products, sign up here. You can also get AI-powered code generation, chat, as well as code reviews directly in the IDE at no cost with the Gemini Code Assist IDE Extension. Footnotes
|
There was a problem hiding this comment.
Code Review
The pull request introduces a new UI component for displaying subagent activity and integrates it into the existing tool result display. It also refactors the subagent invocation logic to stream progress updates as structured objects rather than plain strings, which is a significant improvement for user experience. Tests have been updated to reflect these changes. However, there are a couple of high-severity issues related to redundant truncation logic and a breaking change in the ToolResult API contract, which also touches upon maintaining UI consistency across components.
packages/cli/src/ui/components/messages/SubagentProgressDisplay.tsx
Outdated
Show resolved
Hide resolved
ebd3bbb to
7e25901
Compare
|
Size Change: +12.4 kB (+0.05%) Total Size: 25.8 MB
ℹ️ View Unchanged
|
|
Hi there! Thank you for your contribution to Gemini CLI. We really appreciate the time and effort you've put into this pull request. To keep our backlog manageable and ensure we're focusing on current priorities, we are closing pull requests that haven't seen maintainer activity for 30 days. Currently, the team is prioritizing work associated with 🔒 maintainer only or help wanted issues. If you believe this change is still critical, please feel free to comment with updated details. Otherwise, we encourage contributors to focus on open issues labeled as help wanted. Thank you for your understanding! |
abhipatel12
left a comment
There was a problem hiding this comment.
LGTM with one comment about the thought being sticky and a couple of other nits/thoughts for later
|
|
||
| // Create an activity callback to bridge the executor's events to the | ||
| // tool's streaming output. | ||
| const onActivity = (activity: SubagentActivityEvent): void => { |
There was a problem hiding this comment.
nit: Maybe not now, but could be an opportunity to extract some of this into a separate method, etc to make it easier to read this file.
|
|
||
| const INPUT_PREVIEW_MAX_LENGTH = 50; | ||
| const DESCRIPTION_MAX_LENGTH = 200; | ||
| const MAX_RECENT_ACTIVITY = 3; |
There was a problem hiding this comment.
I think for now this probably makes sense but ideally the CLI/TUI layer is completely in charge of what they want to display, etc. The core should really just emit what it's doing, etc.
But with the state of things, this is great since we already do this for the other fields.
|
|
||
| const INPUT_PREVIEW_MAX_LENGTH = 50; | ||
| const DESCRIPTION_MAX_LENGTH = 200; | ||
| const MAX_RECENT_ACTIVITY = 3; |
There was a problem hiding this comment.
From testing this out, wdyt of keeping it as 3 but making sure the top 1 is always a "thought'.
Sometimes, there a long chain of no thought tool calls and it can be harder to gauge what the subagent is doing. With a thought i have an idea of the latest CoT reasoning.
There was a problem hiding this comment.
I'm not seeing any thoughts bubbling through anymore for some reason. In the interest of unblocking this piece, I'm going to check in this PR and follow up with cleanup for thoughts.
Summary
Improves the subagents UX to show the 3 most recent individual tool calls, live, as they happen.
Note that this is an incremental step and lacks some refinements. e.g.: pressing 'no' in the confirmation dialog fully cancels the subagent rather than pausing execution and letting you intervene.
It also doesn't yet support multiple subagents running in parallel.
Fixes #14309
Pre-Merge Checklist