Index
Sequential index of all units across From LLMs to Secure Agents. Follow the main path in order or explore deep-dive architectural specifications.
Pass 1: Functional Agent Architecture
Section titled “Pass 1: Functional Agent Architecture”Give the reader the minimum software-system vocabulary needed to trace an agent. Do not reteach large language models.
Chapter 1
Reader contract and system mapEstablishes the system vocabulary and diagram notation used to trace an agent safely.
Main path
Chapter 2
Data, control, and trust boundariesSeparates information from instructions and shows where a system must reconsider its assumptions.
Main path
Chapter 3
Requests, events, state, and side effectsExplains how a requested action, remembered state, event record, and outside-world result describe different parts of one workflow.
Main path
Chapter 4
Identity, authority, and least privilege primerExplains identity, delegation, authority, and least privilege in multi-actor software workflows.
Main path
Define agentic systems by observable behavior and system structure, not by product labels.
Chapter 1
What is an agentDefines an agent as an autonomous software system combining a reasoning model with tools, environment observations, and goal-directed control loops.
Main path
Chapter 2
The agent loopExplains the internal mechanics of the agent execution loop, detailing how models perceive environment feedback, decide actions, and execute tools across iterative turns.
Main path
Chapter 3
Workflows versus agentsCompares deterministic code-orchestrated workflows with model-directed autonomous agents, establishing clear criteria for when each architectural pattern should be used.
Main path
Chapter 4
Goals, policies, environments, and autonomyDetails how agent goals, operational policies, environment characteristics, and autonomy levels interact to govern agent behavior and safety.
Main path
Chapter 5
Run lifecycle and terminationDefines the complete lifecycle of an agent run from initialization to termination, detailing execution states, pause mechanisms, and multi-layered stopping conditions.
Main path
Teach stable orchestration patterns before component and framework detail.
Chapter 1
Architecture selection criteriaEstablishes a systematic decision framework and trade-off matrix for selecting among deterministic workflows, single-agent loops, and multi-agent coordination patterns based on latency, cost, determinism, and failure containment.
Main path
Chapter 2
Single-agent and reactive loopsExplores the internal mechanics, state progression, and failure modes of single-agent ReAct loops, detailing how models interleave reasoning with dynamic tool actions and how host runtimes enforce termination guardrails.
Main path
Chapter 3
Sequential, routing, and parallel workflowsDeep dive into deterministic workflow orchestration topologies including linear prompt chaining, conditional routing, parallel sectioning, and consensus voting, emphasizing error isolation and validation gates.
Main path
Chapter 4
Plan and executeExplores the plan-and-execute architectural pattern, detailing how separating strategic task planning from tactical action execution and dynamic replanning improves reliability on complex long-horizon tasks.
Main path
Chapter 5
Evaluator-optimizer and reflectionExplores the evaluator-optimizer and reflection patterns, detailing how decoupled generator and evaluator models iteratively critique, score, and refine outputs against deterministic tests and semantic rubrics.
Main path
Chapter 6
State machines and event-driven graphsExplores state machines and event-driven graphs for AI agents, detailing typed state schemas, cyclic nodes, conditional edge routing, durable checkpointing, and asynchronous human-in-the-loop interruption.
Main path
Chapter 7
Supervisors, handoffs, and agent-as-toolExplores multi-agent coordination architectures, comparing centralized supervisors (manager-worker), decentralized peer handoffs (swarm), and encapsulated subagents (agent-as-a-tool).
Main path
Chapter 8
Architecture trade-offsCompares orchestration patterns across determinism, latency, token expenditure, observability, failure propagation, and termination guarantees to guide minimal architecture selection.
Main path
Provide the complete functional anatomy of production agentic systems.
Chapter 1
Model roles and selectionExplains model roles, capability profiles, selection dimensions, provider adapters, and cost-latency-quality trade-offs in production agentic systems.
Main path
Chapter 2
Routing, cascades, and fallbacksExplores dynamic model routing, progressive escalation cascades, circuit breaker patterns, and multi-provider fallbacks for high-availability agent architectures.
Main path
Chapter 3
Capability, cost, latency, and reliabilityAnalyzes the multi-dimensional trade-offs between model capability, token economics, latency profiles, and operational reliability in production agent workflows.
Main path
Chapter 4
Routing evaluationExplains quantitative evaluation frameworks, cost-quality Pareto benchmarks, and judge calibration techniques for model routers and cascades.
Main path
Chapter 5
Context sources and precedenceExplains the runtime discipline of context engineering, classifying dynamic prompt sources, trust boundaries, and precedence hierarchies in agent loops.
Main path
Chapter 6
Context budgets, selection, and orderingExplains quantitative token budgeting, knapsack chunk selection algorithms, and attention-anchored positional ordering in agent context windows.
Main path
Chapter 7
History, summaries, and compressionExplains conversation history management, lossy and lossless token compaction, tool trace pruning, and recursive summarization for long-running agents.
Main path
Chapter 8
Provenance and context debuggingExplains token-level lineage tracking, context debugging inspection consoles, OpenTelemetry telemetry standards, and boundary integrity auditing.
Main path
Chapter 9
Reactive and reason-act patternsExplains reactive action selection and interleaved ReAct (Thought-Action-Observation) loops, error recovery, and cycle bounds in autonomous systems.
Main path
Chapter 10
Decomposition and plan-executeExplains multi-step goal decomposition, the two-tier Planner-Executor pattern, subtask lifecycle management, and dynamic replanning protocols.
Main path
Pass 2: Threat Modeling & Security (Upcoming)
Section titled “Pass 2: Threat Modeling & Security (Upcoming)”Detailed threat modeling and defensive architectures build directly upon the functional components established in Pass 1.
06
Threat Model & Attack SurfaceAdversarial prompt injection, indirect context contamination, tool squatting, and privilege escalation.
Upcoming
07
Component-Level DefensesContext firewalls, credential isolation, tool token scoping, sandbox runtimes, and approval gates.
Upcoming
08
Secure Reference ArchitecturesHardened multi-agent topologies, supervisor-worker security boundaries, and gateway mediation.
Upcoming
09
Security Testing & EvaluationAutomated red-teaming harnesses, benchmark suites, perturbation testing, and formal verification gates.
Upcoming