🤖 fix: improve outgoing agent communication cards - #4077
Conversation
Give agent reports and task messages matching cards with readable prose, recipient metadata, accessible expansion, and delivery-aware status. Preserve legacy reports and task navigation, with unit and full-app Storybook coverage. Remove one obsolete Config test fixture field to unblock the existing baseline typecheck failure. --- _Generated with [`mux`](https://github.com/coder/mux) • Model: `coder:openai/gpt-6-astra` • Thinking: `high` • Cost: `$21.44`_ <!-- mux-attribution: model=coder:openai/gpt-6-astra thinking=high costs=21.44 -->
This comment has been minimized.
This comment has been minimized.
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 1b67b461c2
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
This comment has been minimized.
This comment has been minimized.
Validate persisted report and message results before rendering, keep transport errors separate from untrusted delivery fields, restore bounded keyboard-accessible message scrolling, and name blank-title report toggles. Add malformed-history and desktop/phone scrolling tests. --- _Generated with [`mux`](https://github.com/coder/mux) • Model: `coder:openai/gpt-6-astra` • Thinking: `high` • Cost: `$21.44`_ <!-- mux-attribution: model=coder:openai/gpt-6-astra thinking=high costs=21.44 -->
|
@codex security review |
|
Codex Review: Didn't find any major issues. 🎉 Reviewed commit: ℹ️ About Codex in GitHubYour team has set up Codex to review pull requests in this repo. Reviews are triggered when you
If Codex has suggestions, it will comment; otherwise it will react with 👍. Codex can also answer questions or update the PR. Try commenting "@codex address that feedback". |
This comment has been minimized.
This comment has been minimized.
Fix the five existing flat-sidebar tests that failed in CI because their workspace-actions fixtures omitted archivingWorkspaceIds. This only completes the test data; production sidebar behavior is unchanged. --- _Generated with [`mux`](https://github.com/coder/mux) • Model: `coder:openai/gpt-6-astra` • Thinking: `high` • Cost: `$21.44`_ <!-- mux-attribution: model=coder:openai/gpt-6-astra thinking=high costs=21.44 -->
|
@codex review The latest commit only supplies missing archive-state sets in five existing sidebar test fixtures exposed by full CI. All 49 sidebar tests and make static-check pass. The communication-card implementation is unchanged from the previous clean code/security review. Generated with |
|
@codex security review |
This comment has been minimized.
This comment has been minimized.
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: febac7e67c
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
Normalize known hook/UI metadata before strict render-time validation, without mutating the original output. Preserve bare pre-hook blocking errors and add frozen hooked-result regression cases for both cards. --- _Generated with [`mux`](https://github.com/coder/mux) • Model: `coder:openai/gpt-6-astra` • Thinking: `high` • Cost: `$21.44`_ <!-- mux-attribution: model=coder:openai/gpt-6-astra thinking=high costs=21.44 -->
|
@codex security review |
This comment has been minimized.
This comment has been minimized.
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 5a7a36db8e
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
Unwrap supported SDK JSON results before hook normalization and schema validation. Treat absent completed outputs as unavailable rather than sent while preserving pending, executing, failed, and interrupted states. Cover both cards with missing-result, wrapper, hook, and error cases. --- _Generated with [`mux`](https://github.com/coder/mux) • Model: `coder:openai/gpt-6-astra` • Thinking: `high` • Cost: `$21.44`_ <!-- mux-attribution: model=coder:openai/gpt-6-astra thinking=high costs=21.44 -->
|
@codex security review |
This comment has been minimized.
This comment has been minimized.
|
Codex Review: Didn't find any major issues. Swish! Reviewed commit: ℹ️ About Codex in GitHubYour team has set up Codex to review pull requests in this repo. Reviews are triggered when you
If Codex has suggestions, it will comment; otherwise it will react with 👍. Codex can also answer questions or update the PR. Try commenting "@codex address that feedback". |
1 participant
Summary
Make outgoing
agent_reportandtask_send_messageentries feel like agent communication rather than command output, matching the readable incoming-report treatment.The branch also fixes existing main-branch validation failures: removes an obsolete
getSessionDirmock from aPartial<Config>fixture and supplies the missing archive-state set in five flat-sidebar fixtures. These are test-data-only corrections; no production workspace/sidebar behavior changes.Validation
make static-checkpassed (including both TypeScript projects).Visual evidence
Desktop
Phone layout and interaction recording
final-interaction.mp4
Long-message scrolling after review fixes
review-scrolling.mp4
Risk
Presentation-only changes to the two outgoing tools; backend delivery is unchanged. Legacy report fallback, all message delivery outcomes, and navigation behavior have regression coverage.
Generated with
mux• Model:coder:openai/gpt-6-astra• Thinking:high• Cost:$21.44