caoqianming
61ff98bb79
loop: thinking spinner 显示耗时 + 累计 token,每轮回打成本一行
...
spinner 文本由后台 daemon 线程每 100ms 用 status.update() 刷新,显示
'thinking... 1.3s ctx 12,345 tok'。每轮 LLM 返回后追加一行 dim
'[in N out N t Xs]',留痕便于回看本轮成本。全 ASCII。
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-05-07 14:05:44 +08:00
caoqianming
80f9934f8f
loop: 用 rich.Markdown 渲染 assistant 输出
...
之前 console.print(f"...{content}") 只解析 [tag] 标签,不识别 **bold**、表格、列表
等 markdown 语法,LLM 回复会以原文显示。改成单独打印前缀后再 console.print(Markdown(content))。
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-05-07 11:19:46 +08:00
caoqianming
dbb778fe10
Phase 4 + 6: capability probe + task 概念 / state.json
...
- core/probe.py + cli.py probe: basic_chat / parallel_tools /
thinking_mode / long_context 四项实测对账 yaml;不进启动路径
- core/task.py + main.py: workspace/tasks/<id>/{state.json, messages.json},
TaskState 跟 mode/desc/status/tokens/timestamps;build_agent 返 5 元组
- cli.py tasks 子命令 + REPL /status /done /abandon /desc;chat 加
--mode/--desc 选项;移除 legacy workspace/sessions/ 兼容
Phase 5 evals 评估后决定不做:个人工具用 dogfooding 判断模型升级,
probe 已覆盖健康检查;造作 case 没区分度,维护成本不划算。
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-05-06 16:21:17 +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
caoqianming