All 54 articles, sorted alphabetically
ADK + A2A interop
Deep-dive on ADK–A2A integration: wrapping agent trees as A2A servers with agent cards, RemoteA2aAgent as in-tree sub-agents, session-to-task identity…
Read article →Agent-as-tool -- composing agents as callable capabilities
Deep-dive on the agent-as-tool composition pattern: a parent calling a sub-agent wrapped as a tool, call-and-return vs sub-agent transfer, encapsulati…
Read article →ADK Agent Lifecycle
From create to deploy to retire — the moving parts.
Read article →ADK Agent Types
The main ADK agent classes: LlmAgent, SequentialAgent, ParallelAgent, LoopAgent — when to use each.
Read article →The ADK Artifact Service
Deep-dive on ADK's Artifact Service: named, versioned, immutable blobs stored out of band in a pluggable backend (in-memory o…
Read article →ADK caching
Deep-dive on caching in agent systems: prompt caching with cache-aligned assembly, semantic caching with similarity thresholds, tool result caching, c…
Read article →ADK callbacks architecture
Deep-dive on ADK's callback system: before/after hooks for agent, model, and tool; the CallbackContext and ToolContext; the s…
Read article →ADK circuit breaker architecture
Deep-dive on the circuit breaker pattern in ADK agents: closed/open/half-open states, rolling rate-and-volume trip conditions, before/after tool callb…
Read article →ADK code executor -- letting the agent run code safely
Deep-dive on the ADK code executor: the computation need, the code execution tool (agent writes and runs code), the sandbox, result feedback, iteratin…
Read article →ADK Context Compaction: Architecture Deep-Dive
How ADK reconciles unbounded session history with a bounded context window — token budgeting, pinned invariants, the elision-to-summarization ladder, …
Read article →ADK Deployment
How to deploy ADK agents to Cloud Run or Vertex AI Agent Engine.
Read article →ADK deployment architecture
Deep-dive on deploying ADK agents: the adk web → API server → Cloud Run/Agent Engine ladder, swappable session/artifact/memory services, per-agent ser…
Read article →Building an Agent Eval Harness
What to measure and how, beyond demo-mode happy paths.
Read article →ADK Evaluation
How ADK's eval framework runs agents against test cases and measures success.
Read article →ADK evaluation architecture
Deep-dive on evaluating ADK agents: eval cases from recorded sessions, Runner replay with fixtures, tool-trajectory vs response scoring, LLM-as-judge …
Read article →ADK Events
How ADK events represent all agent activity (messages, tool calls, tool responses) for observability.
Read article →ADK guardrail architecture
Deep-dive on agent guardrails in ADK: input screening and spotlighting, deterministic tool gates, output sanitization, untrusted-content quarantine, l…
Read article →ADK Guardrails Pattern
Input and output filtering layers.
Read article →ADK LlmAgent
How LlmAgent uses LLM for decision-making, its config options, and interaction with tools.
Read article →ADK long-running tools
Deep-dive on long-running tools in an agent development kit: a tool that starts async work and returns an operation handle, session pause with durable…
Read article →ADK Memory
How ADK's MemoryService stores cross-session facts for agent personalization.
Read article →ADK memory bank
Deep-dive on ADK long-term memory: session ingestion and fact extraction, scoped vector indexing and on-demand retrieval, consolidation with supersess…
Read article →ADK model fallback architecture
Deep-dive on model fallback in the Agent Development Kit: the swappable model abstraction, a health check that classifies failures into retry/next-tie…
Read article →ADK multi-agent architecture
Deep-dive on composing multi-agent systems in ADK: LLM transfer vs AgentTool, Sequential/Parallel/Loop workflow agents, output_key state contracts, es…
Read article →ADK multimodal
Deep-dive on multimodal agents in ADK: typed content parts, native multimodal model reasoning, artifacts for large media, tool outputs as media, multi…
Read article →ADK observability architecture
Deep-dive on observing ADK agents: the event log as audit ground truth, OTel spans per model and tool call, token/cost attribution and anomaly detecti…
Read article →ADK Observability and Traces
Per-turn cost and quality metrics.
Read article →Google ADK Overview
What Google's Agent Development Kit (Python) is, its primitives, and how it fits into Vertex AI ecosystem.
Read article →ADK planning architecture
Deep-dive on planning in ADK: plan generation and decomposition, ReAct vs plan-then-execute, replanning on divergence, reflection loops, workflow agen…
Read article →ADK plugin system architecture
Deep-dive on the Agent Development Kit plugin system: registering plugins with the Runner to run before/after agent, tool, and model callbacks across …
Read article →ADK RAG grounding
Deep-dive on RAG grounding in ADK: chunking and indexing, hybrid retrieval and reranking, retrieval-as-tool, grounded generation with citations, groun…
Read article →ADK runtime architecture
Deep-dive on the ADK agent runtime: Runner turn loop, LLM and workflow agents, event-sourced sessions and state, function/MCP tool execution, callback…
Read article →ADK Safety Evaluation
Red-team patterns for production agents.
Read article →Sandboxing ADK agents
Deep-dive on securing ADK agents against prompt injection and hostile inputs: layered containment with gVisor/container code executors, before/after t…
Read article →ADK session resumption architecture
Deep-dive on ADK session resumption: the session store and append-only event log, checkpointing on each event and recovering on reconnect, resume toke…
Read article →ADK Session Storage
Where state lives across turns.
Read article →ADK Sessions
How ADK sessions store per-conversation state and how the SessionService abstracts backends.
Read article →ADK sessions, state, and memory
Deep-dive on ADK's state architecture: append-only session events and the state fold, session/user:/app:/temp: scopes, Sessio…
Read article →ADK Skills
What ADK Skills are, why they exist, and how they change the way you package agent capabilities.
Read article →ADK State Management
Conversation, memory, and what to actually persist.
Read article →ADK streaming architecture
Deep-dive on ADK's run_live streaming mode: LiveRequestQueue and event pumps, Gemini Live API sessions, barge-in interruption…
Read article →Tool authentication architecture
Deep-dive on tool authentication in agent frameworks like ADK: how an agent calls authenticated external APIs and MCP servers without ever holding a r…
Read article →ADK Tool Design Patterns
What makes a tool the model will use correctly.
Read article →ADK Tools
How to define tools in ADK Python and expose them to LLM agents.
Read article →ADK tool architecture
Deep-dive on ADK's tool layer: function-tool declaration extraction, ToolContext state and auth, parallel execution and error…
Read article →Building a Voice Agent with ADK
STT LLM TTS pipeline.
Read article →ADK Workflow Agents
How SequentialAgent, ParallelAgent, and LoopAgent orchestrate other agents deterministically.
Read article →ADK workflow agents architecture
Deep-dive on ADK workflow agents: SequentialAgent, ParallelAgent, and LoopAgent as deterministic control-flow primitives wrapping LLM sub-agents, sess…
Read article →