zcdsj
/
zcbot
5
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
394 Commits 4 Branches 0 Tags 39 MiB
Commit Graph

6 Commits

Author SHA1 Message Date
caoqianming daf0db854a feat(procs): 后台进程 bg proc + 对话锁/完成通知 + seedance 续查(bump 0.56.0) 
- shell/run_python 加 background=true(模型判断,>~1min 走后台):进程 detach 独立运行,
  不受工具超时/服务重启/蓝绿切换影响;新 check_process 工具查状态/日志尾部/终止
- 状态协议纯文件(<user_root>/.zcbot_procs/<task>/<proc>/,exit_code 出现即终态,
  无 DB 无队列组件):host=stdlib wrapper detach(限时+杀树+日志截尾 10MB);
  docker=专用容器 zcbot-proc-*(同款硬化+iptables,product=proc 与 sandbox
  reaper/shutdown_all 生命周期解耦);web lifespan 每小时 sweep(7d TTL+孤儿容器)
- web:GET /v1/procs(用户级)+ POST /v1/tasks/{id}/procs/{pid}/kill;前端
  [Background] 工具卡活化(spinner/跳秒/停止,历史重渲恢复)+ 完成 toast(跨 task
  可跳转)+ 对话锁(bgproc 运行期 composer 发送→停止/Enter 拦截,先锁后放消
  刷新时序窗口;润色/语音与 streaming 期一致可用)
- seedance:resume_task_id 续查(轮询超时/中断后取回结果,跳过提交不重复计费),
  fast/pro poll_timeout 统一 1200s;取消/超时文案引导续查
- core/llm.py:litellm timeout 显式化(默认 600s 不变,ZCBOT_LLM_TIMEOUT_S 可调)
- DESIGN 新增 §8.12(含边界:不做 job 链/自动续跑/不引 Celery);RUN/PROGRESS/
  CHANGELOG 同步;scripts/test_bgproc_manual.py 手工验证;unittest 全量 201 过

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
 2026-07-10 16:47:11 +08:00
caoqianming 022eade18d feat(research): paper_server 接口对接 API Key(bump 0.39.2) 
paper_server GET 接口收紧要 Key:paper.py 全部请求带 ?api_key=
(env PAPER_SERVER_API_KEY),认证失败(服务端异常处理器压成
HTTP 400 + err_code)转带指引的 RuntimeError。这把低价值只读 key
是唯一刻意进 sandbox 的凭证:run_python env 过滤器加 allowlist
放行,docker 模式 _sandbox_env() 透传进容器(顺带修掉
PAPER_SERVER_URL 覆盖不进 docker 的缺口)。SKILL/RUN/PROGRESS
同步更新。

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
 2026-07-06 11:15:51 +08:00
caoqianming 69fc2599e3 feat(run_python): 过程脚本约定落 task_dir/scripts/ 
模型显式写文件再 script_path 跑的过程脚本统一进 <task_dir>/scripts/
(可见/持久/可重跑),交付产物仍落 task_dir 根。inline code 匿名片段
维持临时用后即焚(host 系统 temp、docker .zcbot_tmp dotfile,均不动)。
改 agent_builder 系统提示工作目录段 + run_python tool description/参数说明。

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
 2026-06-05 15:52:05 +08:00
caoqianming 1c30a9e54e Reduce chat context token usage 2026-06-04 16:41:14 +08:00
caoqianming 97bcd5ae1e feat(skill): paper_server → research skill (search / get_paper / fetch_pdf) 
skills/research/{SKILL.md, paper.py}: 接内部部署 paper_server 的 3 个 helper,LLM 经 load_skill("research") + run_python 调用。范式选 skill 而非 tool/MCP/裸 httpx/lib —— 频次低且 helper 范式让 API 漂移时改一处。tools/run_python.py 注入 PYTHONPATH=base_dir,让子进程能 `from skills.research.paper import ...` 不必折腾 sys.path。base_url 默 http://paper.xxhhcty.xyz:8080,可 PAPER_SERVER_URL env 覆盖。遗留:paper_server 侧 PaperDetailSerializer 加 abstract 字段由用户重新部署。

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
 2026-05-21 13:00:37 +08:00
caoqianming 3a66849953 Initial import: zcbot personal task agent 
DESIGN.md / PROGRESS.md 落地 Phase 1-3:
- core/: LiteLLM 封装 + ReAct loop + 会话持久化 + Anthropic skill registry
- tools/: read/write/edit/glob/grep/shell/run_python/load_skill (Hybrid 范式)
- skills/coding | proposal: WHY+WHAT 风格 SKILL.md
- skills/ppt: 完整渐进披露 (SKILL + 4 references + 3 scripts)
  · 借鉴 hugohe3/ppt-master 的两阶段 + spec lock 思路
  · MSO_SHAPE 图标体系 + 安全区 + 越界检测
  · 默认商务红主题 (#C00000 / #E15554 / #FFC107)
- config/models/: DeepSeek V4 flash/pro 档案

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
 2026-05-06 11:02:59 +08:00