Course program

Four courses that build up: each one rests on the ones before it. The first two are general — development with AI, and the context the agent sees at the moment it answers; the third and fourth show the same discipline inside a regulated ecosystem and inside a closed perimeter. Every lesson ends with three tests and explanations; the courses themselves are taken in the personal cabinet.

Course 1 · AI-driven PDLC

3 chapters · 19 lessons · available in full

Why you need this course. You already write code with an agent, but the result depends on luck: sometimes it works the first time, sometimes much of it — or all of it — has to be rewritten. The course turns that work into a result you can repeat.

What you will learn. Framing a task and assembling its context, the specification as the first artifact, basic checking of the result, the analysis of risks and guard rails. Then the concept of the lifecycle: from Waterfall to working from intent, the roles of an AI team, agentic patterns. The third chapter is a workshop on two instruments, Claude Code and Cursor: subagents, skills, MCP, the links and the boundaries of the agent’s loop.

What you gain. You stop treating the agent as a clever autocomplete and start seeing exactly where the work needs your gate.

How it is built. Three chapters, 19 lessons, every lesson ends with three tests and explanations. Prerequisites — development experience, Git and the command line.

Open in the cabinet

Course diagram: from intent to deliveryIntent sets the specification the agent works from; task context and guard rails feed the same agent. The result — code and tests — goes to human acceptance: a divergence returns to the specification, an acceptance goes to delivery.Intentwhat and why?SpecificationAgentexecutionTask contextGuard railswhat we do not do?Resultcode and testsHuman acceptanceacceptedDeliverydivergence

Course 2 · Context management for AI agents in team development

3 chapters · 24 lessons · prerequisite — the first course

Why you need this course. The agent answers not your question but everything the context window holds. The course is about what fills that window and who is answerable for filling it.

What you will learn. The five layers of context — permanent, project, task, session and tool: each with its own lifetime, its own owner and its own way of reaching the window. Why the permanent layer is not remembered for good but weakens as the session goes on, and why the window is a capacity the layers compete for. Then the development team’s context system: rules, README files and the decision log, the artifact standard and its owner, the normative layer, implicit sources, quality metrics and the symptoms of degradation. The third chapter is work on a task: planning mode, working from a specification and folding the context, review of the context package, test design, research.

What you gain. You stop fixing the agent’s answers and start fixing its input.

How it is built. Three chapters, 24 lessons, every lesson ends with three tests and explanations. Prerequisite — the first course.

Open in the cabinet

Course diagram: the context layers and the context windowFive layers — permanent, project, task, session and tool — converge on the context window and compete for its capacity. The window yields the agent's answer, then metrics and symptoms of context degradation, then folding and refreshing the context, which returns the window to working order.Permanent layerrules, README, decisionsProject layersubsystems and diagramsTask layerspecification and criteriaSession layerthe exchange so farTool layertools, MCP, subagentsThe context windowa capacity the layers compete forThe agent's answerMetrics and symptomsof context degradationFolding and refreshingthe context

Course 3 · Executable knowledge as the subject of development

3 chapters · 20 lessons · prerequisites — the first two courses

Why you need this course. In a regulated environment the subject of development is not code but knowledge, which is turned into executable code by stated rules. Responsibility for the result is different here, and the habits of an ordinary project carry over at too high a price.

What you will learn. Development in a regulated environment: the normative corpus and global conventions, repositories and the boundaries of authority, developing components with agents, the platform MCP servers, gathering knowledge. Extending an existing domain: a program in the ASL language, the authorization loop, domain research and the research document, the provenance of values, calibrating thresholds, a new consumption surface and the analysis of a real divergence. Standing up a new domain: the boundary between the invariant core and the changeable dialect, a seven-step route and a list of prerequisites that no amount of declaration-writing will create.

What you gain. You understand where an authoritative value comes from, and why the question “who decided this” must have an answer.

How it is built. Three chapters, 20 lessons, every lesson ends with three tests and explanations. Prerequisites — the first two courses.

Open in the cabinet

Course diagram: the path of executable knowledgeDomain research and the repositories with their boundaries of authority feed the program in the ASL language. Together with the normative corpus the program passes the authorization loop and becomes published authoritative knowledge, from which the consumption surface and the provenance of a value branch off.Domain researchthe research documentRepositoriesand boundaries of authorityNormative corpusProgramin the ASL languageAuthorization loopPublishedauthoritative knowledgeConsumption surfaceProvenance of a valuewhere did the threshold come from?

Course 4 · The HealthOS development environment

3 chapters · 20 lessons · published together with the environment

Why you need this course. The preceding courses teach you to work with agents that have you: your terminal, your browser, your prompt at any moment. Here you have only a web interface in a personal cabinet, and the fleet of agents has your package of assignments.

What you will learn. The configuration of the environment: why a closed perimeter is needed, what the configuration is made of, how the gates and topologies work, what the alternatives cost, what the MCP servers give. The foundations of a run: the classes of defect and the protection against them, instruments, access and requests, the lifecycle of a resource and the distribution of norms, the task package, the two-phase run and the verification apparatus. Boundaries and traces: what is specific about managing context here, behaviour without a human, the data boundary, roles and secrets, the observability of a run, the answer budget — and an honest account of what does not exist yet.

How it is built. Three chapters, 20 lessons, every lesson ends with three tests and explanations. Prerequisites — the three preceding courses. The course describes the target state of the environment and is published together with it.

Open in the cabinet

Course diagram: agentic work inside a closed perimeterA human hands a task package to the coordinating agent. Inside the closed perimeter run the plan phase, the plan gate, the execution phase, the verification apparatus and the acceptance gate; the result returns to the human. Observability of the run — traces of operations and the answer budget — spans the whole perimeter.A humanTask packagefor the coordinating agentClosed perimeterPhase 1 · planPlan gatePhase 2 · executionVerification apparatusAcceptance gateObservability of the runtraces of operations and the answer budget
HealthOS Academy materials are instructional. Course progress is stored only in your browser and never leaves your device.