Add experimental quiet mode for compact turn display#288
Open
ewilderj wants to merge 2 commits intoxenodium:mainfrom
Open
Add experimental quiet mode for compact turn display#288ewilderj wants to merge 2 commits intoxenodium:mainfrom
ewilderj wants to merge 2 commits intoxenodium:mainfrom
Conversation
When enabled via (setq agent-shell-quiet-mode t), all thought process and tool call fragments within an agent turn are grouped under a single collapsible wrapper. The wrapper label shows the current thought summary. Expanding reveals individual tool calls and thoughts as indented, nested collapsible sections. Implementation uses a dedicated invisible spec (agent-shell-quiet) to manage group visibility without conflicting with the per-fragment collapse state (invisible t). Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Contributor
Author
|
Here's a before/after comparison BEFORE (and goes on for 2.5 more screenfuls) AFTER |
ewilderj
force-pushed
the
ewj/quiet-mode
branch
from
92491c9 to
1696b31
Compare
Instead of one wrapper per turn, create a new collapsible group each time a thought arrives after tool calls. This gives the user visibility into each phase of a long-running turn: ▶ Exploring the codebase... ▶ Found the issue, now fixing... ▶ Running tests... [response text] Each group is independently expandable to see its tool calls. Consecutive thoughts without intervening tool calls stay in the same group. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
ewilderj
force-pushed
the
ewj/quiet-mode
branch
from
1696b31 to
515b1ff
Compare
Contributor
Author
|
Noting again how experimental this is: it becomes quite vulnerable to the various whims of how each model outputs thinking tokens, and whether it streams or not. Pretty sure I found some upstream bugs for Copilot CLI. Still, am trying to make it do a good job on messy data. |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
This is an EXPERIMENTAL mode to show what it would look like to conceal most of the steps between thoughts. In long running turns this shows only the most important agent thoughts and reduces visual noise considerably.
What it does
When
agent-shell-quiet-modeis enabled (t), all thought process and tool call fragments within a turn are grouped under collapsible sections instead of being shown individually. Each section:When a new thought arrives after tool calls, a new collapsible group starts automatically — so long-running turns show multiple phases:
Implementation
agent-shell-quiet.elwith all quiet mode logicagent-shell-uifragments: a wrapper fragment per phase with child fragments for tool callsagent-shell-ui-toggle-fragment-at-pointUsage
No other configuration needed. Set to
nilto return to normal display.