ai-agent-book 精选快照（<2MB 代码与文档，来自 github.com/bojieli/ai-agent-book）

chapter10/book-translation/validation/real_20260730T042330Z/orchestration/glossary.json

@@ -0,0 +1,86 @@
+[
+  {
+    "en": "AI Agent",
+    "zh": "AI 智能体",
+    "pos": "noun",
+    "context": "A system that autonomously plans, executes, and adjusts actions using an LLM as its reasoning engine, context as its working information, and tools as its action interfaces. Examples include Cursor, Deep Research, Manus, and Doubao."
+  },
+  {
+    "en": "LLM (Large Language Model)",
+    "zh": "大语言模型",
+    "pos": "noun",
+    "context": "The reasoning engine of an AI Agent, responsible for understanding intent, planning, decision-making, and judgment. It is trained via pre-training (world knowledge) and post-training (decision-making strategies)."
+  },
+  {
+    "en": "Context",
+    "zh": "上下文",
+    "pos": "noun",
+    "context": "The working set of information available to an Agent at each decision point, including system prompts, tool definitions, user messages, assistant messages, tool results, and dynamic meta-information (e.g., Agent Status Bar). It determines the ceiling of Agent capability."
+  },
+  {
+    "en": "Tools",
+    "zh": "工具",
+    "pos": "noun",
+    "context": "The action interfaces of an Agent, enabling it to interact with external systems (e.g., APIs, file systems, browsers). Tools are categorized into perception, execution, collaboration, event trigger, and user communication types."
+  },
+  {
+    "en": "ReAct Loop",
+    "zh": "推理-行动循环",
+    "pos": "noun",
+    "context": "The core operational loop of an Agent: Reason → Act (tool call) → Observe (tool result) → Reason → Act. It connects the LLM, context, and tools into a coherent system for task execution."
+  },
+  {
+    "en": "Harness Engineering",
+    "zh": "智能体工程框架",
+    "pos": "noun",
+    "context": "The engineering discipline focused on building reliable Agent systems by designing the infrastructure around the LLM, including context management, tool interfaces, constraint mechanisms (Constrain), verification (Verify), and error recovery (Correct)."
+  },
+  {
+    "en": "Agent Skills",
+    "zh": "智能体技能",
+    "pos": "noun",
+    "context": "Modular, loadable knowledge packages that provide specialized domain guidance to an Agent. Skills use progressive disclosure: metadata (name/description) is loaded upfront, while full content is loaded on demand via a dedicated tool."
+  },
+  {
+    "en": "Agent Status Bar",
+    "zh": "智能体状态栏",
+    "pos": "noun",
+    "context": "A mechanism that injects dynamic meta-information (e.g., task progress, tool call counts, environment state) at the end of the context to help the model track runtime state and make better decisions. It distills implicit states into explicit, directly usable knowledge."
+  },
+  {
+    "en": "KV Cache",
+    "zh": "键值缓存",
+    "pos": "noun",
+    "context": "An optimization in LLM inference that caches the key-value states of processed tokens to avoid redundant computation. It requires the prefix (e.g., system prompt + tool definitions) to remain byte-for-byte unchanged for reuse."
+  },
+  {
+    "en": "Prompt Injection",
+    "zh": "提示注入",
+    "pos": "noun",
+    "context": "A security threat where malicious instructions are embedded in external content (e.g., web pages, documents) to hijack an Agent's behavior. Defenses include source tagging, structured roles, and input sanitization."
+  },
+  {
+    "en": "token",
+    "zh": "词元",
+    "pos": "名词",
+    "context": "编辑部指定术语"
+  },
+  {
+    "en": "prompt",
+    "zh": "提示词",
+    "pos": "名词",
+    "context": "编辑部指定术语"
+  },
+  {
+    "en": "latency",
+    "zh": "时延",
+    "pos": "名词",
+    "context": "编辑部指定术语"
+  },
+  {
+    "en": "embedding",
+    "zh": "嵌入向量",
+    "pos": "名词",
+    "context": "编辑部指定术语"
+  }
+]