今日概念 · 已讲过

日期 (Asia/Kuala_Lumpur) 概念 入门原文
2026-10-28 Agent-ready codebase(agent 就绪代码库) https://factory.ai/news/agent-readiness
2026-10-29 Lethal trifecta(致命三要素)/ 提示注入 https://simonwillison.net/2025/Jun/16/the-lethal-trifecta/
2026-10-30 Trunk-based development(主干开发) https://trunkbaseddevelopment.com/
2026-10-31 Pit of Success(成功之坑) https://blog.codinghorror.com/falling-into-the-pit-of-success/
2026-11-01 Hillclimbing(爬坡迭代)/ held-out 留出集 https://claude.dev/blog/automating-eval-design-and-hillclimbing/
2026-11-02 Ralph loop(拉尔夫循环)/ loop engineering https://ghuntley.com/loop/
2026-11-03 Super-agent(超级 agent)/ Router–Worker 两层架构 https://x.com/DeepSeekHarness/status/2116857257245987236
2026-11-04 Formal verification(形式化验证)/ 模型检查(TLA+) https://dantelearn.com/blog/en/posts/2026/10/pstack-tla-formal-verification
2026-11-05 Spec-driven development(规格驱动开发,SDD) https://www.martinfowler.com/articles/exploring-gen-ai/sdd-3-tools.html
2026-11-06 Deterministic hooks(确定性 hooks)vs optional MCP(可选 MCP) https://semgrep.dev/blog/2026/mcp-vs-hooks-ai-agent-security
2026-11-07 Background agents(后台 / 异步 agent) https://code.claude.com/docs/en/agents
2026-11-08 Agent memory(agent 记忆)/ Dreaming(做梦整理) https://www.anthropic.com/research/components-of-agent-memory
2026-11-09 Instruction hierarchy(指令层级) https://arxiv.org/abs/2404.13208
2026-11-10 Agent harness(agent 外壳 / 运行框架) https://www.mihaileric.com/The-Emperor-Has-No-Clothes/
2026-11-11 Blackboard architecture(黑板架构)/ 共享状态协作 https://claude.dev/blog/safely-scaling-multi-agent-teams-with-fable-5-2/
2026-11-12 Property-based testing(属性测试) https://ghuntley.com/answer/ ;https://hypothesis.works/articles/what-is-property-based-testing/
2026-11-13 ADR(Architecture Decision Record,架构决策记录) https://cognitect.com/blog/2011/11/15/documenting-architecture-decisions
2026-11-14 Dependency cooldown(依赖冷却期) https://cooldowns.dev/
2026-11-15 LLM-as-judge(让模型当裁判)/ 裁判校准(TPR/TNR) https://hamel.dev/blog/posts/llm-judge/
2026-11-16 Code Mode(代码模式)/ code execution with MCP https://www.anthropic.com/engineering/code-execution-with-mcp
2026-11-17 Differential testing(差分测试)/ 旧版本当预言机(oracle) https://www.cs.tufts.edu/comp/150FP/archive/bill-mckeeman/DifferentailTesting.pdf
2026-11-18 Reward hacking(奖励作弊 / 钻评分的空子) https://lilianweng.github.io/posts/2024-11-28-reward-hacking/
2026-11-19 Bottleneck / Theory of Constraints(瓶颈 / 约束理论)· AI 让审查成瓶颈 https://en.wikipedia.org/wiki/Theory_of_constraints ;https://fion.ac/jellyfish.pdf
2026-11-20 Threat modeling(威胁建模)· 四个问题 https://www.threatmodelingmanifesto.org/ ;https://cheatsheetseries.owasp.org/cheatsheets/Threat_Modeling_Cheat_Sheet.html