* feat(plugin): support icon field in metadata.yaml for sidebar icon
* feat(sidebar): add isRawTitle support for non-i18n sidebar titles
* feat(sidebar): add usePluginSidebarItems composable for dynamic plugin WebUI entries
* feat(sidebar): inject plugin WebUI items into sidebar before More group
* refactor(plugin-page): remove header, make iframe full-screen
* feat(sidebar): restore plugin WebUI collapsible group
* fix(plugin-page): prevent iframe from capturing mousemove during sidebar resize
- Use absolute positioning instead of negative margin for full-screen layout
- Zero container padding for plugin page route in FullLayout
- Disable pointer-events on iframe during sidebar drag to avoid event capture
* refactor(sidebar): share plugin state reactively instead of polling
- Replace polling + events with module-level reactive shared state
- useExtensionPage.getExtensions() populates pluginSidebarState
- usePluginSidebarItems uses computed() to derive sidebar menu
- Zero additional API calls, updates instantly on any plugin change
* fix(sidebar): restore initial plugin data fetch on sidebar mount
* feat(sidebar): render plugin icons via MDI SVG CDN instead of subset font
- Plugin icons loaded from https://cdn.jsdelivr.net/npm/@mdi/svg@7/svg/
- Removes subset limitation - plugins can use any MDI icon
- Fallback to subset font class for built-in sidebar items
- Default icon remains mdi-puzzle when plugin doesn't specify one
* fix(sidebar): render plugin icons as inline SVG with currentColor for theme matching
* docs: add plugin sidebar icon documentation
* docs: require mdi- prefix for plugin icon field
* chore: ruff format
* fix: address review feedback
- Add SVG sanitization to prevent XSS via v-html (reject <script>, event handlers)
- Extract MORE_GROUP_KEY shared constant to avoid hardcoded i18n key
- Parallel SVG loading with Promise.all instead of serial
- Pure buildPluginItems, mutate iconSvg in place to avoid redundant rebuilds
- Fallback to default icon SVG when loading fails
- Revert accidental pnpm-lock.yaml changes
* chore: remove webui icon
---------
Co-authored-by: Soulter <905617992@qq.com>
* feat(plugin): pass theme through plugin page asset URLs and initial context
Add theme query parameter propagation through the plugin page asset pipeline
so that the bridge SDK initial context includes isDark.
* feat(plugin): inject data-theme and color-scheme into plugin page HTML
Set data-theme on <html> and add color-scheme meta tag server-side
to prevent flash when entering plugin pages in dark mode.
* feat(webui): add isDark getter to customizer store
* feat(webui): sync theme to plugin iframe via URL param and postMessage
Append theme query parameter to iframe src URL and include isDark
in the postMessage context. Watch uiTheme changes to re-send context.
* feat(bridge): auto-apply data-theme from context in plugin bridge SDK
Set data-theme attribute on document.documentElement when context
includes isDark, enabling live theme switching via postMessage.
* docs: add light/dark theme adaptation guide for plugin pages
Add theme adaptation section to existing plugin-pages docs in both
Chinese and English, covering CSS variables and onContext() usage.
* test: add theme sync tests for plugin page bridge and content
Verify isDark in bridge SDK initial context with various theme params,
and verify data-theme and color-scheme injection in rewritten HTML.
* fix(plugin): use case-insensitive regex for head tag in HTML rewrite
Replace string match for <head> with case-insensitive regex to handle
uppercase tags and tags with attributes, preventing duplicate injection
of color-scheme meta tag.
* refactor(webui): generalize isDark getter to support any dark theme
Replace hardcoded PurpleThemeDark check with suffix-based detection
so all dark theme variants are recognized automatically.
* refactor(plugin): extract theme helpers for HTML rewriting
Extract _get_request_theme and _apply_theme_to_html to eliminate
duplicate theme-parsing logic and isolate HTML mutation. Use case-
insensitive regex for head tag detection to prevent duplicate
injection when tags use mixed casing.
* style: apply ruff formatting to plugin page tests
Wrap long function call lines for consistency with project style.
* fix(plugin): handle existing data-theme and case-sensitive fallback in HTML rewrite
Strip any existing data-theme attribute before adding the new one to
prevent duplicate attributes. Use case-insensitive regex for the
<head> fallback insertion to avoid corrupting <html> tag attributes.
* fix(webui): add null guard to isDark getter
Guard against undefined uiTheme to prevent TypeError when the
theme config has not been initialized.
* fix(webui): centralize theme mapping and preserve hash in plugin page URL
Extract themeParam computed to avoid drift between URL and context.
Include hash fragment in iframe URL to support SPA hash routing.
* fix(webui): address CR feedback - deduplicate color-scheme meta, harden isDark getter, consolidate theme source
- _apply_theme_to_html: strip existing color-scheme meta before injecting to
avoid duplicates; merge data-theme strip+add into single-pass regex callback
- customizer.ts: replace brittle endsWith('Dark') with explicit theme name check
- _rewrite_plugin_page_html: use _get_request_theme() directly instead of
reading theme from extra_query_params
* fix(webui): 简化 isDark 推导、优化查询参数构建、使用暗色主题集合
- isDark 推导简化为 theme == "dark"（None == "dark" 为 False，去掉多余三元表达式）
- _prepare_plugin_page_query_params 改为线性构建，先计算再构建字典
- isDark getter 改用显式 DARK_THEMES Set，替代硬编码单值比较
---------
Co-authored-by: lxfight <lxfight@192.168.5.50>
* feat: future task UI
* fix: update filter label for UMO in English and Chinese locales
* feat: enhance cron job management with delivery target handling and UI improvements
* fix: update session label to indicate optional delivery target
* feat: add tooltip for last run time and error in cron job display
* fix(context): restore turn cap, serialize content parts and tool calls for llm compress, fix AftCompact debug log
Three context-compaction regression fixes after #8226:
1. Restore max_context_length -> enforce_max_turns propagation so
normal turn-based truncation works again.
2. Serialize ContentPart and ToolCall objects into plain dicts in
_message_to_dict so llm_compress no longer fails with JSON
serialization errors.
3. Print _provider_messages (compacted) instead of run_context.messages
(unchanged) in AftCompact debug log; truncate long role lists to
first4,...,last4 to avoid log spam.
Assertions in tests are also hardened to avoid coupling to exact prompt
wording.
* fix(tool_loop_agent_runner): simplify context handling by removing redundant provider messages
* fix(tool_loop_agent_runner): rename context manager variables for clarity
* fix: update context compression to use recent token ratio instead of fixed count
* fix: enhance LLMSummaryCompressor to sanitize contexts and improve message handling
* ruff format
---------
Co-authored-by: Soulter <905617992@qq.com>
* fix(console): use toast for pip-install error display
The pip-install dialog displayed error messages as unstyled inline
<small> text with no color differentiation from success messages.
Replaced with useToast() to show errors as red snackbar, consistent
with the rest of the dashboard.
* fix(console): use i18n for pip-install toast fallback messages
* feat: add TOTP two-factor authentication for dashboard login
* fix: ensure TOTP verification uses UTC for accurate time comparison
* fix: update recovery code validation logic for disabling TOTP
* test: add unit tests for TOTP functionality and recovery code validation
* chore: format
* feat: add trust_proxy_headers switch for auth rate-limit IP source
* feat: make dashboard auth rate-limit configurable via system settings
Add auth_rate_limit config block to dashboard settings with enable
(default: true), average_interval (default: 1.0s), and max_burst
(default: 3) options. The dashboard auth middleware now reads from
config instead of using hardcoded values. The average_interval and
max_burst fields are conditionally shown only when rate limiting is
enabled.
* fix: normalize dashboard client IP from trusted proxy headers
* refactor: encapsulate rate limiter state into registry, add TTL
* feat: show dynamic page title during two-factor verification
* fix: require two-factor verification for protected config saves
* chore: format
* refactor: reorganize TOTP verification UI components for better layout
* refactor: clean up recovery stage UI by removing unused styles and improving label handling
* docs: add TOTP two-factor authentication documentation for WebUI
---------
Co-authored-by: Soulter <905617992@qq.com>
* fix: improve template list config handling
* feat(webui): show template list display item
* feat(webui): allow hiding template list hints
* docs: document template list metadata fields
* fix: support file fields in template list configs
* refactor(ltm): redesign long-term memory with context compaction
- Add raw_records / contexts / summaries data model per group
- Add LLM summary compaction strategy alongside truncation
- Add turn-based (_split_into_rounds) granularity
- Add image caption integration into LTM history
- Add tool_call / tool_result persistence into raw_records
- Add active reply support driven by LTM state
- Improve summary injection prefix with system note and delimiters
- Add info-level logging for summary compaction lifecycle
- Clarify default summary prompt with explicit preserve/drop rules
- Add context_guard for history overflow protection in agent runner
- Add internal agent history compaction in agent_sub_stages
- Add comprehensive LTM unit tests and compaction test suites
* fix(ltm): handle malformed JSON in tool args and clean up lock on session removal
* fix(ltm): guard against duplicate system prompt note injection
* fix(ltm): fall back to user message when internal marker parsing fails
- Treat lines starting with <T:CALL>, <T:RES, or <BOT/ as regular user
messages when their respective parsers return None, instead of silently
dropping them. Defensive guard against malformed internal markers.
* fix(ltm): release session lock during LLM summary generation
* fix(ltm): trim raw_records in handle_message to prevent unbounded growth
* perf(ltm): use len(s) instead of len(s.encode()) in trim loop
Avoid allocating a new bytes object for every string when calculating
buffer size in _trim_raw_records. Character count is sufficient for
the approximate memory cap.
* feat(ltm): make user segment truncation limits configurable
* feat(ltm): pre-fill default LTM summary prompt in config and i18n
* refactor(ltm): hardcode internal segment/trim constants
* refactor(ltm): unify compaction strategy with main agent runner
* feat(ltm): add @mention weight marker for group chat messages
* test: fix test failures from LTM compaction unification
* chore(dashboard): remove obsolete LTM compaction i18n metadata
* chore: shrink codebase
* feat(group-chat): implement group chat context management and related functionality
---------
Co-authored-by: Tsukumi <112180165+Tsukumi233@users.noreply.github.com>
Co-authored-by: zenfun <zenfun510@gmail.com>
Co-authored-by: Soulter <905617992@qq.com>
* fix: plugin name in the marketplace does not match the local plugin
* chore: update test
* Update astrbot/dashboard/routes/plugin.py
Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>
* Update astrbot/dashboard/routes/plugin.py
Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>
* Update astrbot/dashboard/routes/plugin.py
Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>
* chore: reformat code
* Clean up unused variables in useExtensionPage.js
Removed unused sets for installed repositories and names.
* fix: 统一repo匹配逻辑，移除fallback
---------
Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>
* feat(plugin): add pages field to plugin list API response
Include discovered page names for each plugin in the /api/plugin/get
response, so the frontend can determine whether a plugin has a WebUI
without an extra detail request.
* feat(webui): add direct WebUI access button on plugin cards
- Add "open-webui" button on plugin cards when plugin has pages
- Navigate to plugin's first page directly from the card
- Adjust PluginPagePage iframe height for better UX
- Add i18n labels (zh-CN/en-US/ru-RU)
* perf(plugin): use asyncio.gather for concurrent page discovery
Replace sequential await loop with concurrent processing to avoid
blocking on disk I/O when discovering plugin pages.
* fix(webui): disable open plugin UI button when plugin is deactivated
Prevent navigation to a disabled plugin's WebUI page which would
result in an error.
* test: update plugin API tests to match pages field in list response
- test_plugin_get_excludes_scanned_pages: expect pages field now present
- test_plugins: use name-based lookup instead of exact count assertion
---------
Co-authored-by: lxfight <lxfight@192.168.5.50>
PR #8015 added 'Noto Sans' to the Google Fonts link and CJK fallback list,
but the font was placed at the end of $cjk-sans-fallback where browsers
never reach it for Cyrillic text. The global $body-font-family also lacked
'Outfit' entirely, causing Vuetify to use CJK fonts as the primary face.
Changes:
- Remove 'Noto Sans' from the end of $cjk-sans-fallback (it is not a CJK font)
- Add 'Outfit' and 'Noto Sans' to $body-font-family before CJK fallbacks
- Update .Outfit class in _container.scss to match the new stack
This ensures:
- Latin text → Outfit
- Cyrillic text → Noto Sans (loaded by vite-plugin-webfont-dl)
- CJK text → Noto Sans SC / PingFang SC etc.
Fixes follow-up to #8015.
The WebUI only loaded Noto Sans SC (Simplified Chinese), which lacks
Cyrillic glyphs. Russian text fell back to system sans-serif, causing
poor rendering depending on the OS.
Changes:
- Load Noto Sans (regular) from Google Fonts alongside Noto Sans SC
- Add 'Noto Sans' at the END of $cjk-sans-fallback (after CJK fonts)
so Chinese text still renders with system CJK fonts first,
while Cyrillic text falls through to Noto Sans.
This ensures both Chinese and Cyrillic text render correctly.
* feat(lark): implement app registration and bot info retrieval
- Add app registration functionality for Lark and Feishu platforms, including endpoints and request handling.
- Introduce polling mechanism for app registration status.
- Create bot info retrieval functionality to fetch bot details after successful registration.
- Enhance dashboard with new UI components for one-click QR setup and manual setup options.
- Update internationalization files to support new features and actions.
- Add unit tests for app registration endpoint resolution and data handling.
* feat(weixin_oc): add WeChat login registration and QR code handling
- Added advanced settings option in update dialog for better user control.
- Implemented detailed progress tracking for update stages including download size and speed.
- Updated localization files for English, Russian, and Chinese to include new strings for update progress and advanced settings.
- Improved UI for update dialog with better layout and responsiveness.
- Enhanced test coverage for update process including progress tracking.
* fix(webui): enforce 12-char dashboard password policy with backend+frontend validation
* fix(i18n): update password policy hints and validation rules for improved security
* test: adapt dashboard auth fixtures for hashed default password
* fix(security): increase PBKDF2 iterations
* feat(auth): implement secure login challenge and proof verification
* chore: ruff format
* fix(auth): update md5 import syntax for consistency
* feat(dashboard): implement random password generation for empty dashboard password
* feat(auth): enforce plaintext password requirement for legacy MD5 hashes
* fix(i18n): update password hint texts to reflect auto-generated initial passwords
* feat(dashboard): implement password change requirement and reset logic
* feat(auth): implement account setup flow and password change requirements
* feat: Implement legacy password support and upgrade mechanism
- Added `hash_legacy_dashboard_password` function for MD5 hashing of passwords.
- Updated configuration handling to store both PBKDF2 and legacy password hashes.
- Introduced logic to check if password storage has been upgraded and if a password change is required.
- Modified dashboard authentication routes to handle legacy password checks and prompts for upgrades.
- Updated frontend to display warnings for legacy password storage and required upgrades.
- Enhanced tests to cover scenarios for legacy password handling and migration to new storage format.
* fix(logo): update text color styles to use CSS variables for consistency
* feat(dashboard): upgrade password storage and enforce change requirement
* fix(dashboard): update minimum password length from 12 to 10 characters
* fix(dashboard): update minimum password length from 10 to 8 characters
---------
Co-authored-by: RC-CHN <1051989940@qq.com>
* feat: enhance plugin page internationalization
- Updated PluginRoute to read initial context from JWT and set it in the bridge SDK.
- Added methods to retrieve locale and plugin metadata for better i18n support.
- Enhanced pluginI18n utility to resolve page-specific translations and added new functions for page titles and descriptions.
- Modified PluginPagePage and PluginDetailPage to utilize new i18n features for dynamic content rendering.
- Improved documentation for plugin page i18n structure and usage.
- Added tests to verify the correct integration of i18n in plugin pages and context handling.
* fix test
* feat: supports plugin to add skills
* fix tests
* fix: fs tools
* Add tests for plugin skills handling and improve skill management
- Implement test for restricted local member reading plugin skill inventory even if the plugin is inactive.
- Ensure that the skill synchronization process retains built-in skills when local skills are empty, including proper handling of plugin paths.
- Update dashboard tests to verify that plugin details include components when requested.
- Enhance skill metadata enrichment tests to include inactive plugin-provided skills for inventory.
- Add filtering tests for plugin skills based on current configuration, ensuring only allowed plugins are considered and inactive plugins are skipped.
Co-authored-by: Copilot <copilot@github.com>
* fix: handle PPIO platform context-length error messages (#7888)
* fix: 压缩算法删除 user 消息 Bug 修复
* perf: improve truncate algo
* fix: improve context length error detection for PPIO platform compatibility
- Extend error detection to handle PPIO's error message format:
'The input is longer than the model's context length'
- Add case-insensitive matching using .lower() for robustness
- Maintain backward compatibility with existing 'maximum context length' check
This fixes the issue where PPIO platform models (e.g., ppio/zai-org/glm-5-turbo)
would fail with AgentState.ERROR due to unrecognized context length errors.
---------
Co-authored-by: Soulter <905617992@qq.com>
* fix: 支持微信客服文件消息 (#7923)
* fix: 支持微信客服文件消息
* fix: remove WeCom file message placeholder
* fix(provider): fix Anthropic custom headers and system prompt compatibility (#7587)
* fix(provider): fix Anthropic custom headers and system prompt compatibility
- Pass custom_headers via AsyncAnthropic's `default_headers` parameter
instead of creating a separate httpx.AsyncClient. This avoids
`isinstance` check failures when multiple httpx installations exist
on sys.path (e.g. bundled Python + system Python).
- Use list format for the `system` parameter (`[{"type": "text", ...}]`)
instead of a plain string. The list format is supported by the official
Anthropic API and is also compatible with third-party API proxies that
reject the string format.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* fix(provider): fix Anthropic custom headers and system prompt compatibility
- Pass custom_headers via AsyncAnthropic's `default_headers` parameter
instead of creating a separate httpx.AsyncClient. This avoids
`isinstance` check failures when multiple httpx installations exist
on sys.path (e.g. bundled Python + system Python).
- Use list format for the `system` parameter (`[{"type": "text", ...}]`)
instead of a plain string. The list format is supported by the official
Anthropic API and is also compatible with third-party API proxies that
reject the string format.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* Add test unit
---------
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
* perf: improve logic of adding models
Co-authored-by: piexian <piexian@users.noreply.github.com>
* chore: remove redundant logger messages and improve log clarity
Co-authored-by: Copilot <copilot@github.com>
* chore: ruff format
* docs: update knowledge base docs
closes: #7962
* fix(#7907): send_message_to_user cron 场景下 session 容错 (#7911)
* fix: send_message_to_user cron 场景下 session 容错 (#7907)
- LLM 在主动场景可能只传 session_id 而非完整三段式，
from_str 失败时用 current_session 补全前两段。
Co-authored-by: Copilot <copilot@github.com>
* fix: 限制 session 补全仅对裸 session_id 生效，避免误修带冒号的错误输入 (#7907)
* feat: add session information to cron job payload
Co-authored-by: Copilot <copilot@github.com>
* fix: improve clarity and consistency of safety mode prompts
Co-authored-by: Copilot <copilot@github.com>
---------
Co-authored-by: Copilot <copilot@github.com>
Co-authored-by: Weilong Liao <37870767+Soulter@users.noreply.github.com>
Co-authored-by: Soulter <905617992@qq.com>
* perf: tool rendering in conversation page (#7937)
* fix(dashboard): route conversation history tool messages through ToolCallCard
When viewing conversation history, large tool outputs (e.g. a single
git log --stat producing tens of KB) caused the browser renderer to
freeze. Root cause: formattedMessages mapped every role (including
tool / system / _checkpoint) into user/bot bubbles, and bot plain
strings went through markstream-vue's MarkdownRender. Single 88KB
tool messages plus 88-of-them adding up to ~349KB of synchronous
markdown parsing was enough to block the main thread for 5+ seconds.
This patch:
- Indexes tool-role messages by tool_call_id
- Filters formattedMessages to user/assistant only — tool, system and
_checkpoint roles no longer render as standalone bubbles
- Converts assistant.tool_calls (OpenAI shape, with tc.name/tc.arguments
fallbacks) into the existing tool_call MessagePart, attaching the
paired result so MessageList's ToolCallCard renders it (default
collapsed, no longer feeds large strings into the markdown renderer)
- Drops empty placeholder plain parts when an assistant message only
carries tool_calls
- Sets ts/finished_ts to 0 as a sentinel: ToolCallCard.toolCallDuration
returns "" when startTime <= 0, suppressing a misleading "0ms"
duration that would otherwise appear because conversation history
has no real timing data
Behavior change: tool results are now embedded in their assistant's
ToolCallCard.result instead of appearing as separate bot bubbles.
This matches the main chat UI's behavior.
Fixes#7929
Refs #7372#7456
* style(dashboard): use single scrollbar in conversation history preview
ToolCallCard's result/args panes have their own max-height + overflow,
which produced a nested scrollbar when nested inside the history
preview's already-scrollable .conversation-messages-container. Override
those constraints inside the preview only — the outer 500px-bounded
container already provides scroll bounds, so a single scrollbar feels
cleaner. The main chat UI is unaffected.
---------
Co-authored-by: wanger <wanger@example.com>
* fix: ruff format
* feat: add python tool timeout param (#7953)
* feat: add python tool timeout param
* Update python.py
---------
Co-authored-by: Weilong Liao <37870767+Soulter@users.noreply.github.com>
* fix: 钉钉连接超时后自动重连失败 (#7924)
* fix: improve DingTalk adapter error handling in run() method
* fix: add retry logic for DingTalk SDK task unexpected exit
* fix: use task.add_done_callback to wake thread on task completion, handle UnboundLocalError
* refactor: extract retry logic into handle_retry helper function
---------
Co-authored-by: Blueteemo <Blueteemo@users.noreply.github.com>
---------
Co-authored-by: Copilot <copilot@github.com>
Co-authored-by: leonforcode <leonbeyourside01@gmail.com>
Co-authored-by: AstralSolipsism <134063164+AstralSolipsism@users.noreply.github.com>
Co-authored-by: Pink YuDeer <wer00001@outlook.com>
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
Co-authored-by: piexian <piexian@users.noreply.github.com>
Co-authored-by: NayukiMeko <ChibaNayuki@163.com>
Co-authored-by: wanger <122891289+10knamesmore@users.noreply.github.com>
Co-authored-by: wanger <wanger@example.com>
Co-authored-by: Haoran Xu <3230105281@zju.edu.cn>
Co-authored-by: 千岚之夏 <108566281+Blueteemo@users.noreply.github.com>
Co-authored-by: Blueteemo <Blueteemo@users.noreply.github.com>
* fix(dashboard): route conversation history tool messages through ToolCallCard
When viewing conversation history, large tool outputs (e.g. a single
git log --stat producing tens of KB) caused the browser renderer to
freeze. Root cause: formattedMessages mapped every role (including
tool / system / _checkpoint) into user/bot bubbles, and bot plain
strings went through markstream-vue's MarkdownRender. Single 88KB
tool messages plus 88-of-them adding up to ~349KB of synchronous
markdown parsing was enough to block the main thread for 5+ seconds.
This patch:
- Indexes tool-role messages by tool_call_id
- Filters formattedMessages to user/assistant only — tool, system and
_checkpoint roles no longer render as standalone bubbles
- Converts assistant.tool_calls (OpenAI shape, with tc.name/tc.arguments
fallbacks) into the existing tool_call MessagePart, attaching the
paired result so MessageList's ToolCallCard renders it (default
collapsed, no longer feeds large strings into the markdown renderer)
- Drops empty placeholder plain parts when an assistant message only
carries tool_calls
- Sets ts/finished_ts to 0 as a sentinel: ToolCallCard.toolCallDuration
returns "" when startTime <= 0, suppressing a misleading "0ms"
duration that would otherwise appear because conversation history
has no real timing data
Behavior change: tool results are now embedded in their assistant's
ToolCallCard.result instead of appearing as separate bot bubbles.
This matches the main chat UI's behavior.
Fixes#7929
Refs #7372#7456
* style(dashboard): use single scrollbar in conversation history preview
ToolCallCard's result/args panes have their own max-height + overflow,
which produced a nested scrollbar when nested inside the history
preview's already-scrollable .conversation-messages-container. Override
those constraints inside the preview only — the outer 500px-bounded
container already provides scroll bounds, so a single scrollbar feels
cleaner. The main chat UI is unaffected.
---------
Co-authored-by: wanger <wanger@example.com>
