A thread you can test
Agent Engineering
24 notes move from the word to a real choice at work — understand it first, then decide whether to use it.
Each note stands alone, or becomes the next step in this thread.
THE QUESTION THIS PAGE ANSWERS
ANSWER FIRSTWhat is Agent Engineering, and which AI decisions does it change?
Four capabilities: Plan / Tool / Memory / Act — click to see real-world cases This page keeps the related concepts, common mistakes, and practical notes in one reading thread.
First decide whether you are blocked by a definition, a choice, or verification; then choose the closest of the 24 notes below.
Start with “Agent: AI That Gets Things Done,” then restate the conclusion using your own task.
Do not treat every method in a topic as interchangeable. The answer changes with the input, risk, and acceptance bar.
THIS QUESTION THREAD
Put the word back inside the choice it changes.
Agent: AI That Gets Things Done
Four capabilities: Plan / Tool / Memory / Act — click to see real-world cases
The Secret of Tool Calling
Model outputs JSON → framework parses & executes → result injected back; four-step flow visualization
5 Messages Behind One Conversation
Deconstructing the real Function Calling chain: the user sees 1 reply, but 5 API messages happen behind the scenes
The Art of Tool Descriptions
Same functionality, but good vs bad descriptions differ by 3× in success rate — a contrast experiment
Multi-Tool Orchestration: Parallel vs Serial
isConcurrencySafe determines whether tools can run in parallel; scheduling strategy visualization
MCP Protocol: The USB Port for Tools
Comparing stdio / SSE / Streamable HTTP transport methods; data-flow animation
ReAct in Practice: Weather Query Full Chain
Thought / Action / Observation step-by-step demo; complete 7-step Agent chain
Short-Term Memory = Context Window
Message list visualization showing each message's Token usage and role label
Context Compression: Four Defense Lines
60% trim → 75% micro-compress → 85% fold → 95% emergency; drag the slider to watch the process
Long-Term Memory: Vector Retrieval
Embedding → vector database → semantic search; design decisions for topK and minScore
From Embeddings to Milvus
Semantic similarity, ANN, and the responsibility boundary of a vector database
The Milvus Mental Model
Collection, Schema, Entity, Index, Search, Query, and Load
Hands-on Milvus
Connect, create, batch insert, index, search, query, and delete
From Retrieval to RAG
Chunking, filtering, hybrid search, RRF, reranking, and evaluation
ReAct Loop: Think → Act → Observe
A task to remove console.log goes through 14 iterations, including self-correction
5 Patterns of Agent Deadlock
Parameter format errors / hallucinated tools / infinite recursion / insufficient info / API exceptions
Permissions & Security
5 permission modes + LLM risk levels + Human-in-the-loop design
Skill: Helping Agents Avoid Detours
Skill = workflow instructions + tool-calling guidance; a "laundry on the balcony" analogy for good loops vs bad loops
The Essence of Skill
Good loop vs bad loop upgraded; how Skill changes an Agent's execution path
Scaffolding: From Prototype to Product
Simulate an Agent booking flights & hotels; full comparison with and without scaffolding; 5 key capabilities explained
5 Engineering Guardrails
Iteration cap, output truncation, timeout control, interrupt recovery, context emergency
Multi-Agent Collaboration
Sub-Agent dispatch, worker threads, parallel vs serial execution strategies
Observability
Event stream visualization, Token tracking, OpenTelemetry integration
Agent Engineering Overview
From four capabilities to production deployment — the complete Agent knowledge map on one page