* fix(agent): improve send_message_to_user tool description to prevent misuse
Fixes#6402
The AI was inappropriately using send_message_to_user tool in normal
conversations for text replies, causing duplicate messages (once via tool,
once via normal response).
Root cause: Tool description was not clear enough about when to use it.
Changes:
- Restructure description with clear sections using markdown
- Emphasize two valid use cases:
1. Sending media files (image, record, video, file) in any conversation
2. Proactive messaging scenarios (cron jobs, background tasks)
- Add explicit warning: Do NOT use for normal text replies
- Explain consequence: Using for text causes duplicate messages
This approach (better documentation) is safer than restricting tool
registration, which would break media file sending in normal conversations.
Alternative to PR #6413 which was closed due to breaking media delivery.
* fix: correct syntax error and use triple-quoted string for description
Address review feedback:
- P0: Add missing closing parenthesis (was causing SyntaxError on import)
- Use triple-quoted string for better readability (suggested by gemini-code-assist)
- Remove explicit newline characters, let Python handle line breaks
---------
Co-authored-by: ccsang <ccsang@users.noreply.github.com>
- tool_loop_agent_runner.py: add wrapper coroutine for create_task
- tool_image_cache.py: add cast for Self return type
- astr_agent_tool_exec.py: add type ignore for add_tool
- astr_main_agent.py: add type ignores for tool operations
- astr_main_agent_resources.py: add type ignore for msg_dict
- astrbot_config_mgr.py: add type annotations
- MCPTool_T must be imported at runtime, not in TYPE_CHECKING
- Add __future__ annotations for forward references
- Fix remaining conflict markers in openai_source.py
* fix: send SSE heartbeat to prevent WebChat disconnect during compression
When context compression triggers with slow reasoning models (e.g.
deepseek-reasoner), the backend can go 30+ seconds without pushing any
SSE data. The client-side EventSource / browser then assumes the
connection is dead and disconnects, causing the WebUI to hang
indefinitely since it never receives the eventual response.
Fix: yield an SSE comment (`: heartbeat`) on every empty poll cycle.
Comment lines are defined in the SSE spec as keep-alive signals --
the EventSource API ignores them but the HTTP connection stays open.
Fixes#6938
* chore(dashboard): extract SSE heartbeat to constant
---------
Co-authored-by: Yufeng He <40085740+universeplayer@users.noreply.github.com>
Co-authored-by: RC-CHN <1051989940@qq.com>
* feat: add new StatsPage for enhanced statistics overview
- Introduced StatsPage.vue to provide a comprehensive overview of statistics with various metrics and charts.
- Implemented fetching and displaying of base and model provider statistics.
- Added unit tests for provider statistics persistence in the database.
* style: refine card styles and remove unnecessary shadow for improved aesthetics
- fix(openai_source): remove dead code comment
- fix(common.ts): use resolveApiUrl helper for live-log endpoint
- style(mdi-subset): clean up icon CSS
- feat: add translation check script and unit tests
Introduce session-level state management for tools that need to maintain
state across conversation turns within the same session (UMO).
- Add ToolSessionManager class for central per-(UMO, tool_name) state
- Add ToolSessionState with set_persistent(key) support
- Add is_stateful flag to FunctionTool base class
- Wire session_manager through run_context in all agent runners
- Update ExecuteShellTool to use framework session manager
- Update CLAUDE.md with stateful tool documentation
BREAKING: Tools that were manually managing session state via _sessions
dict should migrate to use the framework's ToolSessionManager for proper
lifecycle management and persistence support.
* fix(gsvi_tts): Use the correct calling method (#5638)
Add some configuration items for GSVI
* fix(gsvi_tts): add default value for api_key in provider configuration
* fix(gsvi_tts): Adjust wherever the Authorization header is built to only include it when `self.api_key` is truthy
Delete some comments
* chore: ruff format
---------
Co-authored-by: Soulter <905617992@qq.com>
Resolved conflicts:
- openai_source.py: keep dev version with abort_signal filtering
- customizer.ts: keep dev version with viewMode functionality
- useSessions.ts: keep dev version with pendingSessionId handling
- platformUtils.js: keep dev version with correct tutorial links
- AddNewPlatform.vue: keep dev version with correct docs link
- FullLayout.vue: keep dev version with viewMode-based logic
- VerticalHeader.vue: keep dev version with viewMode-based logic
`abort_signal` (asyncio.Event) is passed via **kwargs into payloads during
tool_call streaming, causing "Object of type Event is not JSON serializable"
when the OpenAI client tries to serialize the request body.
Regression test added: test_prepare_chat_payload_strips_non_json_serializable_kwargs
* fix(wecom): fallback to regular message API when kf API returns 40096
When sending WeCom messages via kf/send_msg, if the API returns error
40096 (invalid external userid), fall back to the regular message/send
API. This handles internal employees who don't have external userids.
Fixes the issue where internal WeCom users (e.g. WangCong) would cause
kf API to fail with 'invalid external userid' error.
* fix(wecom): improve error handling for kf API fallback to regular message API
---------
Co-authored-by: Soulter <905617992@qq.com>
