What does this PR do?

Fixes Desktop thinking disclosure auto-open state so each reasoning disclosure only checks its own reasoning group for pending content.

Before this change, ReasoningAccordionGroup scanned from the current group start through the rest of the assistant message. If a later thinking block was still running, older completed thinking blocks in the same message could be treated as pending and reopen automatically.

This bounds the pending check to startIndex..endIndex, matching the existing content check and the actual reasoning group rendered by that disclosure.

Related Issue

Discord support thread: https://discord.com/channels/1053877538025386074/1514062044263874704

Type of Change

• 🐛 Bug fix (non-breaking change that fixes an issue)
• ✨ New feature (non-breaking change that adds functionality)
• 🔒 Security fix
• 📝 Documentation update
• ✅ Tests (adding or improving test coverage)
• ♻️ Refactor (no behavior change)
• 🎯 New skill (bundled or hub)

Changes Made

• apps/desktop/src/components/assistant-ui/thread.tsx: bound pending detection for ReasoningAccordionGroup to the current reasoning group.
• apps/desktop/src/components/assistant-ui/streaming.test.tsx: added a regression covering separated reasoning groups where the earlier completed group stays collapsed while a later group is running.

How to Test

1. Run npm run test:ui --workspace apps/desktop -- src/components/assistant-ui/streaming.test.tsx -t "does not reopen an earlier completed thinking group".
2. Confirm the regression passes: first disclosure aria-expanded=false, second disclosure aria-expanded=true.
3. Optional manual check: stream an assistant message with multiple non-consecutive thinking blocks in Desktop and confirm older completed thinking blocks do not reopen when a later thinking block is added.

Focused validation run locally:

• ✅ npm run test:ui --workspace apps/desktop -- src/components/assistant-ui/streaming.test.tsx -t "does not reopen an earlier completed thinking group"

Note: the full streaming.test.tsx file currently has an unrelated existing failure in renders an incomplete streaming reasoning fenced code block as a code card, which also fails when run by itself in this checkout.

Checklist

Code

• I've read the Contributing Guide
• My commit messages follow Conventional Commits (fix(scope):, feat(scope):, etc.)
• I searched for existing PRs to make sure this isn't a duplicate
• My PR contains only changes related to this fix/feature (no unrelated commits)
• I've run pytest tests/ -q and all tests pass
• I've added tests for my changes (required for bug fixes, strongly encouraged for features)
• I've tested on my platform: WSL/Linux running Desktop Vitest target

Documentation & Housekeeping

• I've updated relevant documentation (README, docs/, docstrings) — or N/A
• I've updated cli-config.yaml.example if I added/changed config keys — or N/A
• I've updated CONTRIBUTING.md or AGENTS.md if I changed architecture or workflows — or N/A
• I've considered cross-platform impact (Windows, macOS) per the compatibility guide — or N/A
• I've updated tool descriptions/schemas if I changed tool behavior — or N/A

For New Skills

N/A

Screenshots / Logs

User report shows Desktop Client v0.16.0 (+21) and Backend v0.16.0 (+21) with prior thinking blocks expanded again after a later thinking block appears.