Obsidian Vault Feed

81 Markdown notes, excluding .trash.

RSS feed

1. Post by @dillon_mulroy on X
   workflow/agent/Post by @dillon_mulroy on X.md
2. fix-bug
   prompts/fix-bug.md
3. The Problem: Agents in Silos
   Clippings/Cross-Repo Handoff Protocol Coordinating AI Coding Agents Across Separate Repositories.md
4. command，没有 / 前缀，表达动作
   Clippings/从 github cli 学习如何设计 agent friendly cli.md
5. What I'll cover
   Clippings/How's Linear so fast A technical breakdown.md
6. Just enough inference with evals
   Clippings/How to use evals and prompt engineering to ship a local model that matches frontier performance.md
7. MMD 渲染配方 —— Unity 实现可行性分析
   art/MMD渲染配方-Unity可行性分析.md
8. 个人向2d风格渲染分享
   art/个人向2d风格渲染分享.md
9. 先给结论
   workflow/project/unity工程优化建议.md
10. 基于证据复核后的 Unity 文档改进建议
    workflow/project/unity工程优化建议2.md
11. what you'll get from this article
    Clippings/How to Become a Hermes Agent Operator.md
12. 0. 先问最危险的问题
    workflow/agent/这个 harness 真的需要存在吗？.md
13. 1. 把 coding agent 看成控制系统
    workflow/agent/把钱学森《工程控制论》的思想映射到 coding agent harness.md
14. 0. 调研后的关键判断
    workflow/agent/Agent 主导的 coding agent harness 架构.md
15. Why HTML?
    workflow/Using Claude Code The Unreasonable Effectiveness of HTML.md
16. 一、Skills Library 的上限，不是能力不足，而是工程失稳
    workflow/agent/从 Skills 到分层 Workflow：AI Agent 工程化的下一层抽象.md
17. 未命名
    chat/未命名.md
18. v3.0：针对"AI 主力重构 + 多项目 + 你 review"的最小可用方案
    chat/AI-Native 项目文档体系.md
19. 0x00
    workflow/agent/0x00.md
20. 你的理解 8.5/10，方向对，但有两个隐藏陷阱
    chat/ARCHITECTURE和AGENTS.md
21. 未命名
    workflow/ux/未命名.md
22. The same file can help one task and hurt another by 30%
    Clippings/A good AGENTS.md is a model upgrade. A bad one is worse than no docs at all.md
23. 重构 Chiron ContextEngine
    workflow/agent/上下文/重构 Chiron ContextEngine.md
24. 未命名
    workflow/agent/上下文/未命名.md
25. 基础架构
    workflow/agent/Codex 桌面应用技术栈全景分析.md
26. The Container Bottleneck
    Clippings/Building a Virtual Filesystem for Mintlify's AI Assistant.md
27. The Mistake: Why First Drafts Fall Short
    workflow/ui/设计循环：让Codex成为UIUX的魔术师【完整指南】.md
28. 利用器工程：为编码代理准备 Unity C# 代码库
    workflow/project/利用器工程：为编码代理准备Unity CSharp代码库.md
29. 架构一览：五件套都在干嘛
    Clippings/我用 Cloudflare 免费搭了一套 AI 内容流水线，真的能跑起来.md
30. 背景
    Clippings/用 Claude Code 将三万行 Go 项目移植到 Rust：Agent Team 实践与 Harness 效率优化.md
31. Make the repository legible to agents
    workflow/project/利用器工程：为编码代理准备TypeScript代码库.md
32. 为什么一定要有工作流？
    workflow/ui/很多人问我的 AI UI 工作流，被我完整拆成一条完整链路：先讲需求和目标，再落 UI Spec，再做视觉探索，再进入组件化实现，最后再修复.md
33. UI 的问题，往往不是代码问题
    workflow/ui/别再只让 Codex 写代码了，它更适合接管整条 UI 生产线.md
34. 为什么不要直接把图转成代码？
    workflow/ui/不要再直接把 UI 图转成代码了，先看这份 UI Spec 模板.md
35. Abstract
    workflow/agent/Agentic Harness Engineering Observability-Driven Automatic Evolution of Coding-Agent Harnesses.md
36. Knowledge
    Clippings/Google Antigravity Documentation 6.md
37. Browser Recordings
    Clippings/Google Antigravity Documentation 5.md
38. Screenshots
    Clippings/Google Antigravity Documentation 4.md
39. Walkthrough
    Clippings/Google Antigravity Documentation 3.md
40. Implementation Plan
    Clippings/Google Antigravity Documentation 2.md
41. Task List
    Clippings/Google Antigravity Documentation 1.md
42. Google Antigravity Documentation
    Clippings/Google Antigravity Documentation.md
43. The long-running agent problem
    inbox/Effective harnesses for long-running agents.md
44. 1\. Harness Design: Split by Role, Coordinate via Filesystem
    inbox/Claude Code 源码蒸馏 - Harness Engineering 实践记录.md
45. Some Practical Multi-agent Experiments
    inbox/Multi-Agents What's Actually Working.md
46. 一、Tool Calling 已经走到了一个十字路口
    inbox/REPL 是 MCP 和 CLI 的最终形态.md
47. Agent
    agent/README.md
48. Archive
    archive/README.md
49. Civic
    civic/README.md
50. Companies
    companies/README.md
51. Concepts
    concepts/README.md
52. Deals
    deals/README.md
53. Hiring
    hiring/README.md
54. Household
    household/README.md
55. Ideas
    ideas/README.md
56. Inbox
    inbox/README.md
57. Media
    media/README.md
58. Meetings
    meetings/README.md
59. Org
    org/README.md
60. People
    people/README.md
61. Personal
    personal/README.md
62. Programs
    programs/README.md
63. Projects
    projects/README.md
64. Prompts
    prompts/README.md
65. Sources
    sources/README.md
66. Writing
    writing/README.md
67. Computational vs Inferential
    inbox/harness-engineering-for-coding-agent-users.md
68. # 暗工厂日记 #1：组织理论、误差数学与自主软件工厂的架构
    sources/vadim-comanescu/dark-factory-diaries-1.md
69. # 黑暗工厂日记 #2：别再组建机器人团队了
    sources/vadim-comanescu/dark-factory-diaries-2.md
70. 碟子里的弹簧
    sources/vadim-comanescu/dark-factory-diaries-3.md
71. 读取问题
    sources/vadim-comanescu/dark-factory-diaries-4.md
72. 黑暗工厂日记 #5：每个上下文窗口一个任务
    sources/vadim-comanescu/dark-factory-diaries-5.md
73. 黑暗工厂日记 #6：为什么选择Pi作为编码代理
    sources/vadim-comanescu/dark-factory-diaries-6-pi.md
74. 灯火通明
    sources/vadim-comanescu/the-lights-are-on-everywhere.md
75. # 黑暗工厂日记 #2：别再组建机器人团队了
    sources/vadim-comanescu/dark-factory-diaries-2-停止构建机器人团队.md
76. # 暗工厂日记 #1：组织理论、误差数学与自主软件工厂的架构
    sources/vadim-comanescu/dark-factory-diaries-1-自主软件工厂的架构.md
77. 灯火通明
    sources/vadim-comanescu/the-lights-are-on-everywhere-灯火通明.md
78. 黑暗工厂日记 #6：为什么选择Pi作为编码代理
    sources/vadim-comanescu/dark-factory-diaries-6-为什么选择Pi作为编码代理.md
79. 读取问题
    sources/vadim-comanescu/dark-factory-diaries-4-管理单次代理运行中的上下文.md
80. 黑暗工厂日记 #5：每个上下文窗口一个任务
    sources/vadim-comanescu/dark-factory-diaries-5-每个上下文窗口一个任务.md
81. 碟子里的弹簧
    sources/vadim-comanescu/dark-factory-diaries-3-让工厂防错.md