Skip to content

v2.10.1

Latest
Latest

Choose a tag to compare

View all tags
@github-actions github-actions released this 17 Apr 20:46
v2.10.1
05d891e
This commit was signed with the committer’s verified signature.
dalamudx Dalamud
GPG key ID: C77139B06E5CF161
Verified
Learn about vigilant mode.

What's Changed

  • feat(agent): upgrade langchain to v1.0+ in ea4e0dd
  • feat(agent): upgrade langchain to v1.0+ in 5cded77
  • fix: qqbot wechatbot 模块循环依赖问题 in e4451c7
  • fix agent bug in 5739ca7
  • feat: Enhance agent persona to be cute, playful, and anthropomorphic with increased emoji usage and updated Chinese examples. in 7e9c489
  • feat: Introduce MemoryMiddleware and PatchToolCallsMiddleware to the agent, and add EditFileTool and WriteFileTool for file manipulation. in 30b332a
  • fix：Agent会话时间调整为1天，通过自动压缩控制上下文 in c13abfd
  • 更新 version.py in 210c323
  • fix: bug (jxxghp#5605) in 4873ffd
  • Fix agent memory (jxxghp#5607) in 41f53d3
  • add LLMToolSelectorMiddleware in 565e10b
  • refactor: Update agent stream processing to support 'v2' chunk format and prevent emitting empty content. in 9b00a5f
  • feat(agent): Telegram与Agent相互时支持流式输出 in 4bc67dc
  • feat(agent): Telegram与Agent相互时支持流式输出 in c85805b
  • 更新 base.py in 6ca6a94
  • 更新 version.py in d4a15d3
  • 更新 init.py in 873e383
  • 更新 init.py in faa046e
  • add TAVILY_API_KEYS in 17519d5
  • 更新 base.py in 5744228
  • 更新 init.py in 6f4df91
  • 更新 init.py in 40c79b2
  • fix: 修复集缩略图文件名错误，episode-thumb-xx 改为 视频文件名-thumb.xx in 630d13a
  • 更新 version.py in f2daa63
  • feat: 新增 LLM_MAX_TOOLS 配置项，支持按需启用 LLMToolSelectorMiddleware in 309b7b8
  • feat(agent): support skills in e82494c
  • feat(agent): support skills in 80e35fa
  • rollback aiopathlib in 78b8b30
  • fix bug in be42c78
  • docs: Add guidelines for creating new skills, including directory structure and SKILL.md format. in 2406438
  • fix: change U115_APP_ID in a66e082
  • 更新 skills.py in 5d3511c
  • feat: 新增修改下载任务Agent工具，查询下载任务支持返回标签 in aae5000
  • feat: 查询下载任务工具支持按标签过滤 in 1c16b8b
  • fix: 修复查询下载任务工具访问TransferTorrent不存在字段的问题 in cfc8d26
  • 更新 base.py in 3c7fdfe
  • fix: 优化Agent消息排版 in 17373bc
  • feat: 新增Agent浏览器操作工具(browse_webpage)，支持通过Playwright控制浏览器进行网页交互 in 1249929
  • 更新 version.py in 4a78929
  • feat: QuerySitesTool 增加返回 cookie 字段 in dcca318
  • fix: 修复技能加载时误读目录导致 IsADirectoryError 的问题 in d94b596
  • fix: 修复Agent技能元数据中Path对象无法msgpack序列化的问题 in 94ed065
  • fix: 修复Agent流式输出时回复消息未记录到数据库的问题 in e6e50d7
  • ci: 优化发布脚本，自动生成分类更新日志 in 04b8475
  • fix: build actions in 10807a6
  • feat: 文件名为辅助中文标签时使用父目录标题识别 in de4dbf2
  • feat: 智能体增加定时任务(Jobs)管理和心跳唤醒机制 in 94c75eb
  • fix: 修复 channel 为空时系统提示词 markdown_spec 占位符未替换导致 KeyError in 940cece
  • feat: 后台任务(定时唤醒)跳过流式输出，仅广播最终结果 in 0399ab7
  • fix: 后台任务使用非流式执行，仅发送模型最后一条回复 in 0a4091d
  • fix: 优化心跳提示词，后台任务只生成执行结果摘要 in d38b6df
  • fix: AI_AGENT_JOB_INTERVAL 默认为0 in e598d5e
  • 更新 version.py in c822782
  • feat: 同一会话消息排队顺序处理，不同会话互不影响 in a6f16dc
  • feat: 流式输出消息超长时自动分段发送，消息长度限制纳入渠道能力管理 in 4b4b808
  • feat: 新增插件相关agent工具（查询插件、查询插件能力、运行插件命令） in a1fa469
  • feat: 新增 moviepilot-api 技能，支持全量 REST API 调用；技能中间件自动同步内置技能到用户目录 in d5d6442
  • feat: 新增 ActivityLogMiddleware，自动记录每次交互的活动日志并注入系统提示词 in 83bf59d
  • refactor: 活动日志摘要改用 LLM 总结替代文本截取 in c85d3ad
  • feat: tmdbid优先识别，同ID电影/电视剧通过元数据自动消歧 in 1acf783
  • feat: 记忆为空时自动引导用户设置偏好；优化默认回复风格为简约直接 in 213a298
  • style: 调整默认回复风格，简洁干练但保留适度的俏皮和emoji in a12147d
  • fix: agent过滤模型思考/推理内容，不输出thinking到用户 in 7d0c790
  • fix in 740b0a1
  • 更新 version.py in 20387a0
  • 优化记忆结构 in e7d1469
  • 更新 memory.py in 27cc559
  • feat(agent): 支持Slack和Discord渠道的流式输出功能 in d7c62fb
  • 新增agent删除下载历史记录工具 in a25c709
  • fix(agent): 解决内存文件读取编码问题 in d8c8606
  • 更新 version.py in c3e96ae
  • feat(agent): 新增AI_AGENT_VERBOSE开关，控制工具调用过程回复及提示词输出 in 8d7ff2b
  • v2.9.21 in 75d4c01
  • fix verbose_spec in aaeb4d2
  • fix current_date in f859d99
  • refactor: Invert AI agent verbose mode condition and strengthen silence instructions for tool calls. in a8ecdc8
  • refactor: Update agent package initialization and imports. in 6806900
  • feat(agent): 在系统提示词中注入MoviePilot配置信息 in 8e0a9f9
  • refactor(prompt): 优化MoviePilot系统信息注入，统一日期与环境信息展示 in c3b7a50
  • refactor: 移除异步pop操作 in 296d815
  • refactor: 优化工具消息发送逻辑 in bfea017
  • feat(agent): 添加reset方法支持流式消息原地更新 in f70eaf9
  • 优化Agent提示词：日期改为当前时间，注入系统安装目录，强调简洁回复 in a977948
  • 更新 version.py in 8c603ba
  • feat(agent): 添加数据库信息到系统提示词 in b9b19c9
  • feat(agent): 添加PostgreSQL用户名到数据库信息 in cdf40a7
  • feat(telegram): 保持正在输入状态直到消息发送完成 in a1a72df
  • agent提示词中注入PostgreSQL数据库密码 in e3fee39
  • feat(agent): 支持图片消息处理 in b349aa2
  • refactor: 移除多余的局部导入 in 2d358e3
  • feat(agent): 增加工具执行权限控制 in 02439f5
  • fix(agent): 修复 MessageChannel.QQBot 不存在的错误 in ca9cbc1
  • feat(agent): 为需要管理员权限的工具添加 require_admin 字段 in 0cab21b
  • fix: 修复QQ渠道名称为qq in 4bba7db
  • fix: 修复QQ渠道key映射 in 7bd7751
  • 去除集图片的-thumb后缀，使图片名称与视频文件名称一致 in 8999c3a
  • feat(agent): 接入Exa API用于网络搜索 in 7010963
  • feat(telegram): 优化Telegram文件下载与base64转换逻辑，重构消息发送相关代码 in 9c51f73
  • agent工具增加管理员权限校验：查询站点、查询已安装插件、查询插件能力、查询站点用户数据、刮削元数据 in cbff2fe
  • refactor: 移除多余的空行，优化类型导入及方法为静态方法 in e61299f
  • 更新 version.py in 2e034f7
  • docs: 新增SKILL.md，补充MoviePilot重启与升级操作说明 in fc47382
  • fix: 后台模式不发送工具调用消息 in 84b938c
  • fix(telegram): 转发频道消息无法接收及内容丢失 in 321bf94
  • feat(telegram): 支持 disable_web_page_preview 禁用链接预览 in 5440dba
  • fix(telegram): 修复 disable_web_page_preview 传递给不支持的方法及 UTF-16 偏移量问题 in d9eb329
  • refactor(telegram): 显式传递 disable_web_page_preview 参数避免 @Retry 下修改 kwargs in c68450f
  • fix(system): 补充 fuse 挂载关键词 in c9149d1
  • 更新 version.py in c4cfd70
  • feat: add NotificationType for Agent messages in 1f526ad
  • fix: 智能体唤醒后消息发送问题 in 244112b
  • fix: 智能体唤醒后消息发送问题 in ec0915e
  • fix(meta): 修复首括号被误删导致标题识别错误 in 388afa8
  • feat: TransferRename 事件增加 source_path 源文件路径参数 in 60ac901
  • fix: change logger warning to debug for empty Discord configs in 4cda182
  • feat: 整理失败时AI智能体自动重试 in d7a5c32
  • 更新 version.py in 376c653
  • feat: TransferRename 事件增加 source_item 源文件信息 in f70881b
  • fix: 修复获取Google模型列表阻塞事件循环及缺少代理配置的问题 in 974d5f2
  • fix: 修复Gemini 2.5思考模型工具调用时thought_signature缺失导致400错误 in 5b892b3
  • 更新 init.py in 8960620
  • refactor: SendMessageTool message_type 改为消息标题 in 663e37b
  • 删除 .DS_Store in b777e8c
  • fix(helper): 统一redis缓存键 in 2f151ce
  • 修复 QQbot渠道依旧会重复发送消息问题 in 8b9dc0e
  • feat(subscribe): 优化洗版订阅合集的识别 in 2ae98d6
  • test: 补充媒体刮削路径、图片与事件流程测试 in eb5e17a
  • feat(agent): add command-execute skill for intelligent command dispatch in 7806267
  • refactor(agent): rename run_plugin_command to run_slash_command to avoid confusion with execute_command (shell) in a8fb4a6
  • refactor(agent): rename list_all_commands to list_slash_commands and skill to command-dispatch in 9fefd80
  • fix(agent): resolve circular import by lazy-importing Command in run_slash_command and list_slash_commands in 902394f
  • feat(agent): add database-operation skill for SQL access with auto SQLite/PostgreSQL detection in 8f38c06
  • 更新 SKILL.md in 0ecbcb8
  • feat(agent): add /stop_agent command for emergency stop of agent reasoning in 8109ffb
  • refactor(agent): optimize database-operation skill to read DB info from system prompt <system_info> in 02cb5df
  • feat(agent): add tools for querying and updating custom identifiers in ac9c959
  • fix(agent): 基于langgraph_step过滤中间步骤思考文本，抽离ThinkTagStripper类 in 1e338e4
  • 更新 version.py in ad40b99
  • 更新 query_subscribes.py in 0132919
  • 更新 query_subscribe_history.py in c0c08b0
  • 更新 query_installed_plugins.py in 7072d2c
  • 更新 list_directory.py in 79ec3ed
  • 更新 search_media.py in 9c75c2d
  • 更新 search_person.py in 1ca54af
  • 更新 search_web.py in 6b01901
  • agent工具支持翻页及取消数量限制 in 5acfd68
  • 查询订阅历史工具有名称过滤时不分页直接返回所有匹配结果 in 3c7cd21
  • feat(transfer): 新增 TransferOverwriteCheck 事件支持插件介入覆盖判断 in 4d9f17b
  • feat(transfer): TransferOverwriteCheck 支持插件提供源文件真实大小 in 1b489ba
  • feat(agent): 合并同批次整理失败的agent重试调用，避免重复浪费token in 0c51d79
  • feat(skills): 内置技能支持版本号管理，更新时自动覆盖旧版本 in dc428e7
  • feat(plugin): 聚合插件侧栏导航 in fd6c9d5
  • 更新 version.py in abc4154
  • feat(agent): 优化Agent流式输出与工具消息发送逻辑 in 770065d
  • fix(agent): 简化非流式模式下工具调用的消息处理逻辑 in d5ed9bc
  • fix(agent): 非流式模式下啰嗦模式仍需发送工具调用中间消息 in 78e8faa
  • fix(agent): 非流式模式下不发送任何工具中间消息 in 9675d19
  • 更新 init.py in 18906e5
  • fix(agent): start_streaming 始终标记流式状态以支持 buffer 收集 in aafc4b3
  • fix(agent): 流式+啰嗦模式下渠道不支持编辑时立即发送工具消息 in 22f3244
  • refactor(agent): 重命名 can_edit_message 为 is_auto_flushing 更贴切语义 in 75fca97
  • feat(plugin): reuse plugin wheels in batch dependency install in 351de8b
  • fix(plugin): avoid caching failed plugin index responses in 8cd12ab
  • refactor: 移除 agent 批量重试逻辑中的多余 try 块并优化缩进 in bf12a86
  • feat(search): 支持渐进式（SSE）搜索资源并实时返回搜索进度与结果 in 10543ee
  • 更新 version.py in 8082558
  • fix(http): close non-success responses safely in 217fcfd
  • fix(http): use explicit success checks in async callers in 1af5f02
  • fix(http): require 200 for share reporting requests in df3294c
  • fix(http): close helper responses consistently in 815d83b
  • fix(telegram): pass disable_web_page_preview through edit_message_text in a4f2c57
  • Add agent image support for Telegram and Slack in 6c5fae5
  • Fix forwarded image payload parsing for agent channels in edf3946
  • Add tracing logs for agent image message flow in cb32365
  • Fix Telegram agent image download path in bf2d2cb
  • Expand image and edit support across messaging channels in 2f53fd3
  • fix(subscribe): refresh total episodes before completion check in 1ceb889
  • fix(subscribe): reset tv episode counts in history response in 61c695b
  • chore(subscribe): update last_update when refreshing episode totals in 5d07381
  • 更新 version.py in 9d8bd50
  • 更新 version.py in 9dababb
  • feat(agent): add voice message support with TTS/STT for Telegram and WeChat in e5f97cd
  • fix(system): expose backend dev flag only in dev mode in 8d938c2
  • feat(agent): add audio message extraction and download support for Slack, QQ, Discord, SynologyChat, and VoceChat in e09f9ad
  • fix jxxghp#5661 插件package文件不存在时不报错 in f1f4743
  • fix(plugin): handle 404 plugin index and None response safely in ae5207f
  • fix(subtitle): create missing download root before saving subtitles in 1688063
  • fix(transfer): clean migrated queue jobs in d36dd69
  • fix(transfer): fail stale queue tasks on errors in 6aec326
  • refactor(transfer): make queue job migration explicit in eda73e1
  • fix(transfer): tighten queue cleanup edge cases in 8182894
  • feat(agent): support file attachments and local file replies in 7a5e513
  • feat: expose AI agent flag in user global settings in 2ae2ea8
  • 更新 llm.py in 1176723
  • 更新 version.py in bf127d6
  • Improve agent image capability routing in 13c3c08
  • 更新 config.py in 65a4b74
  • Add manual AI redo flow in 04c2a1e
  • add full-stack local cli install flow in 9a2a241
  • fix local cli install and config workflow in 02a98f8
  • fix bootstrap script for macos bash in e9c8db9
  • update reused bootstrap repo before setup in 262bd68
  • run setup config step inside venv in 49b50e5
  • fix local db initialization model registration in 60996be
  • extend setup wizard for database and agent in 5995b3f
  • bump version to v2.10.0 in 6fcc5c8
  • support local restart and site auth wizard in 61a0a66
  • fix local update dirty check in 70350aa
  • automate local bootstrap prerequisites in 89f6164
  • bump version to v2.10.1 in 77b95d1
  • feat: add retry actions for failed transfers in 011535f
  • feat: add agent button choice workflow in cc31c66
  • refactor: generalize agent interaction requests in e0e21e3
  • relax local install python requirement to 3.11 in 810cb0a
  • feat(cli): optimize installation command and support initializing user password in eea7e3b
  • 优化资源包下载逻辑，只下载对应操作系统和Python版本的sites文件 in f3d9f25
  • fix: 修正 docker 和 update.sh 中 python_version 的格式以匹配 sites.cpython-* 命名规则 in df1b206
  • fix: 修复子进程环境下获取事件循环失败的问题 in 787db8f
  • feat: add download state manager and minor changes in 751944a

Full Changelog: v2.9.15...v2.10.1

Contributors

Retry
Assets 2
Loading