Skip to content

Media upload for upstream PR 13056 - #78

Closed
Bil0000 wants to merge 2765 commits into
mainfrom
show-subagent-model-effort
Closed

Bil0000 wants to merge 2765 commits into
mainfrom
show-subagent-model-effort

Conversation

@Bil0000

@Bil0000 Bil0000 commented Sep 22, 2026 •

Copy link
Copy Markdown
Owner

Temporary image upload for pingdotgg#13056. This draft will be closed after the images are linked upstream.

Before: subagent model without effort

After: subagent model with saved effort

Before: cropped subagent hover card

After: cropped subagent hover card

juliusmarminge and others added 30 commits September 20, 2026 22:32
Keep collapsed model controls in a strip, contain transition overflow, and preserve timeline spacing. Render approval requests as regular grouped worklog entries.

Implemented with GPT-6-Astra via Codex.
Adapt question attachments and Android push verification to V2 requests and shell events. Preserve composer transitions and compact worklogs while integrating upstream loading, navigation, and mobile changes. Release consumed application replay pages without retaining earlier batches.
Accumulate owned step usage once and preserve partial or unavailable telemetry for failed, interrupted, or reconnected turns.
…dotgg#10880)

Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
Co-authored-by: Claude Fable 5 <noreply@anthropic.com>
Co-authored-by: Mike Olson <mwolson@member.fsf.org>
Co-authored-by: Julius Marminge <julius0216@outlook.com>
Co-authored-by: Julius Marminge <julius0216@outlook.com>
Bil0000 and others added 24 commits September 21, 2026 12:09
Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>
Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
…g#12747)

Co-authored-by: Julius Marminge <51714798+juliusmarminge@users.noreply.github.com>
Co-authored-by: Claude Fable 5 <noreply@anthropic.com>
…g#12846)

Co-authored-by: Julius Marminge <51714798+juliusmarminge@users.noreply.github.com>
Co-authored-by: Claude Fable 5 <noreply@anthropic.com>
…gg#12790)

Co-authored-by: Julius Marminge <51714798+juliusmarminge@users.noreply.github.com>
Co-authored-by: Claude Fable 5 <noreply@anthropic.com>
…otgg#12409)

Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
Co-authored-by: Claude Fable 5.1 <noreply@anthropic.com>
Co-authored-by: Julius Marminge <51714798+juliusmarminge@users.noreply.github.com>
…tgg#11870)

Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
Co-authored-by: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com>
Co-authored-by: Julius Marminge <51714798+juliusmarminge@users.noreply.github.com>
Co-authored-by: Claude Fable 5 <noreply@anthropic.com>
…#12839)

Co-authored-by: Claude Fable 5.1 <noreply@anthropic.com>
Co-authored-by: Julius Marminge <51714798+juliusmarminge@users.noreply.github.com>
@github-actions github-actions Bot added the vouch:trusted PR author is trusted by repo permissions or the VOUCHED list. label Sep 22, 2026
@Bil0000 Bil0000 closed this Sep 22, 2026
@Bil0000

Bil0000 commented Oct 3, 2026

Copy link
Copy Markdown
Owner Author

Current-main evidence from the real app with stopped local fixtures. No provider turn ran.

The hover card uses the composer speed labels and filled lightning icons beside model effort. Standard and Normal have no bolt. Fast has one bolt. Ultrafast has two bolts. Unreported or unsupported speed stays hidden.

The live GPT-6.1 Sol catalog offers Standard and Fast. Ultrafast and Flex below use isolated catalog fixtures for existing code paths. The real account catalog was not changed.

Before:
Before

Standard:
Standard

Fast:
Fast

Ultrafast catalog fixture:
Ultrafast

Claude Normal:
Normal

Claude Fast:
Claude Fast

No speed reported:
No speed

Flex catalog fixture:
Flex

Captured by GPT-6.1 Sol, High, in the native Codex harness through T3 Code.

speed-before

speed-after-standard

speed-after-fast

speed-after-ultrafast

speed-after-normal

speed-after-claude-fast

speed-after-unreported

speed-after-flex

@Bil0000

Bil0000 commented Oct 3, 2026

Copy link
Copy Markdown
Owner Author

Current-main evidence from the real web app and native iOS development client. This uses stopped local projection fixtures. No provider ran and no private conversation data was used.

The latest report appears above the composer. Its timeline row starts collapsed. Native MCP Tools expands and scrolls to the final mcp__capture__tool_60 row. Custom Agents expands. The native card stays bounded above the input.

Native before:
Native before

Native after:
Native after

Native long-section scroll:
Native long section

Native section opening and scrolling, shown at 3x playback speed:
context-native-evidence.mp4

Web before:
Web before

Web after:
Web after

Captured by GPT-6.1 Sol, High, in the native Codex harness through T3 Code.

context-native-before

context-native-after

context-native-long

context-native-evidence.mp4

context-before

context-after

@Bil0000

Bil0000 commented Oct 3, 2026

Copy link
Copy Markdown
Owner Author

Current-main before/after images and interaction video from the real web app with a stopped sample conversation.

At 1512 x 948 with the sidebar open, the workspace card is 312 pixels wide. Real pointer gestures request preview widths of 320, 480, and 900 pixels; the last fits at 845 pixels. The composer and timeline rectangles stay exactly unchanged during resize, close, and reopen. At 390 x 844 the inline card is hidden and page scroll width is 390 pixels.

The integrated check found and fixed one compatibility issue: when the card folded into popover presentation, its closed popover state removed the saved inline chat lane. The saved inline setting now keeps that lane fixed. The focused regression test passes.

The preview uses the real app frame, stores, resize handles, and close control with synthetic preview state. This headless web capture does not prove native Electron browser page execution; its preview content can show a reconnecting message.

Before, same laptop size:
Before

After:
After

Large floating preview:
Preview

Narrow layout:
Narrow

Open, resize, close, reopen, and narrow layout:
layout-evidence.mp4

Captured by GPT-6.1 Sol, High, in the native Codex harness through T3 Code.

layout-before

layout-after

layout-after-preview

layout-after-narrow

layout-evidence.mp4

@Bil0000

Bil0000 commented Oct 3, 2026

Copy link
Copy Markdown
Owner Author

Updated icon-only subagent hover-card evidence for upstream pingdotgg#13056. Fast uses one filled bolt before effort; Ultrafast uses two. Standard/Normal, Flex, and unreported speed have no speed text or icon. All captures use stopped local app fixtures. Ultrafast and Flex use an isolated catalog fixture; no provider turn ran.

speed-icons-before

speed-icons-after-standard

speed-icons-after-fast

speed-icons-after-ultrafast

speed-icons-after-normal

speed-icons-after-claude-fast

speed-icons-after-unreported

speed-icons-after-flex

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

vouch:trusted PR author is trusted by repo permissions or the VOUCHED list.

Projects

None yet

Development

Successfully merging this pull request may close these issues.