Find the chapter behind the question.
96 curriculum documents. Search directly, browse by area, or follow the factory lifecycle.
96 of 96 documents
Start Here
12 documentsAI Software Factory and Mission Control
Establish the factory's governing idea and separate it from a coding assistant, coding agent, or generic agent platform.
Intent · Plan · Execute · Verify · Deliver · Learn · Factory SystemCanonical Glossary
This glossary defines terms by the responsibility they own. A useful definition also states what the concept does not prove or authorize.
Intent · Plan · Execute · Verify · Deliver · Learn · Factory SystemPlatform Blueprint and Operating Playbook
This overview connects the factory's product thesis, capability model, reference architecture, reliability and security posture, learning system, adoption model, and success measures. It is a scope map, not a claim that every capability is
Intent · Plan · Execute · Verify · Deliver · Learn · Factory SystemIntent-to-Delivery Lifecycle
Turn the factory's memorable value stream into precise records, decisions, controls, and evidence.
Intent · Plan · Execute · Verify · Deliver · Learn · Factory SystemSoftware Factory Stack Boundaries
Give every component one clear responsibility and replacement boundary.
Intent · Plan · Execute · Verify · Deliver · Learn · Factory SystemReading Paths
The curriculum is a reference system, not a book everyone must read in the same order. Choose the path that matches the decision you need to make. Every path uses the same canonical architecture and vocabulary.
Intent · Plan · Execute · Verify · Deliver · Learn · Factory SystemAI Software Factory Topic Index
Use this map when you have a topic in mind but do not know where it belongs in the curriculum. “Start here” gives the shortest useful orientation. “Deep dive” supplies the production architecture. “Practice” points to a lab or exercise wher
Intent · Plan · Execute · Verify · Deliver · Learn · Factory SystemCapability Coverage and Maturity
This page prevents curriculum breadth, editorial maturity, and implementation proof from being mistaken for one another. A subject may be well explained and not implemented. A mechanism may exist in code and lack accepted operating evidence
Intent · Plan · Execute · Verify · Deliver · Learn · Factory SystemExternal Reviewer Guide
Thank you for reviewing AI Software Factory Mastery. The most valuable review identifies an incorrect boundary, missing failure mode, unsupported claim, unclear explanation, or exercise that cannot produce the evidence it promises.
Verify · Learn · Factory SystemCurriculum Changelog
Material changes to scope, terminology, maturity, architecture, and review status are recorded here. Minor punctuation and formatting corrections may be grouped.
Learn · Factory SystemDetailed Architecture Coverage Matrix
Give every material factory responsibility one canonical name, owner, chapter, control boundary, and validation path.
Intent · Plan · Execute · Verify · Deliver · Learn · Factory SystemStart Here
Intent → Plan → Define Agent → Execute through Harness → Apply Skills → Evaluate → Improve → Deliver Software
Intent · Plan · Execute · Verify · Deliver · Learn · Factory SystemVision
1 documentFirst Principles
1 documentOperating Model
6 documentsThe Human-Agent Operating Model
Define how people and agents divide decisions, execution, oversight, and accountability.
Intent · Plan · Verify · Learn · Governance · Human AuthorityFactory Economics and Operating Metrics
Agent activity is easy to measure and easy to mistake for value. Tokens, sessions, generated lines, tool calls, and pull request volume can all rise while customer outcomes slow, defects increase, and engineers spend more time recovering or
Intent · Plan · Verify · Learn · Governance · Human AuthorityGoverned Continuous Learning and Recursive Improvement
A factory that never learns repeats failures and requires permanent manual tuning. A factory that changes its prompts, policies, workflows, evaluations, or authority automatically can become unpredictable. Continuous learning must improve t
Intent · Plan · Verify · Learn · Governance · Human AuthorityEnterprise Adoption and Factory Maturity Model
Enterprises often buy agent capability faster than they build the operating system required to govern it. A maturity model must describe observable organizational capability, not enthusiasm, model intelligence, or the percentage of code gen
Intent · Plan · Verify · Learn · Governance · Human AuthorityCompounding Engineering and Human Attention
Teams repeatedly correct agents for the same repository convention, testing requirement, architectural boundary, review preference, or failure mode. If those corrections remain inside individual conversations, the organization pays for the
Intent · Plan · Verify · Learn · Governance · Human AuthorityEnterprise Governance Operating Model and Decision Rights
Problem solved: Make authority explicit from enterprise risk appetite to one system, release, incident, and autonomy decision. Three levels: Executive governance; enablement and control; accountable system and business ownership. Critical s
Intent · Plan · Execute · Verify · Deliver · Learn · Governance · Human AuthorityDomain Model
5 documentsThe Authoritative Delivery Hierarchy
Preserve intent, authority, causality, evidence, and acceptance across every delivery record.
Intent · Plan · Control Plane · ContractsFactory Configuration, Workflow Contracts, and Execution Manifests
An approved WorkOrder states what may be achieved. It does not fully describe the machinery that will perform the work. If workflow, agents, prompts, tools, models, repository paths, context, budgets, and recovery rules are resolved after d
Intent · Plan · Control Plane · ContractsSpecification Engineering, Executable Requirements, and Plan Assurance
An agent cannot reliably implement intent that has never been made precise. “Improve onboarding” may be a valid business objective, but it does not identify the users, desired behavior, constraints, unacceptable outcomes, or proof needed fo
Intent · Plan · Control Plane · ContractsMulti-Repository Development and Coordinated Delivery
A business outcome may require coordinated changes to an API, client, shared schema, infrastructure, deployment configuration, and documentation stored in different repositories. An agent working from one checkout can miss a dependent contr
Intent · Plan · Control Plane · ContractsFactory System Inventory, Classification, and Lifecycle
Maintain one accountability record for each material autonomous delivery system without copying its subordinate registries.
Intent · Plan · Execute · Verify · Deliver · Learn · Control Plane · ContractsAgent Factory
4 documentsAgent Factory Capability Supply Chain and Registries
Explain how reusable agent capabilities move from an idea to an approved, discoverable, operable factory component.
Define · Execute · Verify · Learn · Agent Factory · Capability SupplyCapability Packaging, Versioning, and Dependency Resolution
Make agent capabilities reproducible, composable, and safely replaceable.
Define · Execute · Agent Factory · Capability SupplyCapability Evaluation, Certification, Promotion, and Retirement
Define how a capability earns, retains, loses, and exits production eligibility.
Verify · Learn · Agent Factory · Capability SupplyTool, Skill, and Integration Contract Reference
Make every callable capability discoverable, authorizable, testable, observable, recoverable, and revocable.
Define · Execute · Verify · Learn · Agent Factory · Capability SupplyRuntime Architecture
9 documentsControl Plane and Execution Plane
Separate durable authority and policy from long running, failure prone execution.
Execute · Runtime · Harness · EnvironmentRuntime Orchestration and State Machines
Approving a WorkOrder does not execute it. Between authorization and a review ready result lies a distributed process that may run for minutes or hours, cross process and provider boundaries, survive restarts, wait for human decisions, and
Execute · Runtime · Harness · EnvironmentTasks, Attempts, Leases, Idempotency, and Recovery
Agent execution fails in ambiguous ways. A worker can disappear while its process continues. A retry can duplicate a commit or pull request. A stale worker can report success after another worker has taken over. If a Task and an execution t
Execute · Runtime · Harness · EnvironmentSandboxed Execution, Isolation, and Publication Boundaries
An implementation agent executes untrusted generated actions against valuable source code. A worktree prevents branch collision, but it does not contain process, credential, network, resource, or host risk. Conversely, a disposable virtual
Execute · Runtime · Harness · EnvironmentFactory Observability and Agent Runtime Telemetry
An agentic factory can be busy while making no progress, cheap while producing unsafe work, or successful according to a model while violating authority. Operators need to understand outcome, control state, execution health, cost, and evide
Execute · Runtime · Harness · EnvironmentAI Software Factory Reference Architecture
Connect intent, authority, execution, verification, delivery, operations, and learning through one canonical architecture.
Execute · Runtime · Harness · EnvironmentDevelopment Environments, Compute, and Composable Infrastructure
An agent cannot build or test software merely because a model and repository are available. It needs the correct compilers, dependencies, services, identities, network paths, test data, browser or device capabilities, and preview surfaces.
Execute · Runtime · Harness · EnvironmentCoding Harnesses, Adapters, and Agent Protocols
Coding harnesses expose different tools, permission models, session formats, hooks, subagents, context behavior, output events, sandboxes, and completion semantics. A factory that shells out to a CLI may appear provider neutral while silent
Execute · Runtime · Harness · EnvironmentOrchestration Component Model and Runtime Contracts
Define the twelve component families that turn authorized work into bounded, durable, observable execution.
Plan · Execute · Verify · Learn · Runtime · Harness · EnvironmentAI Engineering
11 documentsAI Systems Foundations for Software Factory Architects
Explain only the AI concepts that change factory architecture, cost, security, evaluation, or operations decisions.
Intent · Plan · Execute · Verify · Learn · Agent Runtime · Models · ContextAgent Architecture, MCP, Tools, Context, and Memory
Define the complete, versioned runtime surrounding a model driven engineering worker.
Plan · Execute · Verify · Learn · Agent Runtime · Models · ContextModel Routing, Evaluations, and Capability Selection
No model is best for every factory operation. Strong models cost more and may be slower. Fast models may lack tool use, context, reliability, or risk approval. Provider outages and rate limits make a single route fragile. Choosing solely by
Plan · Execute · Verify · Learn · Agent Runtime · Models · ContextData, Knowledge, Context, and Semantic Engineering
Separate data fitness, reusable knowledge, semantic meaning, and task specific context so failures can be located and corrected.
Plan · Execute · Verify · Learn · Agent Runtime · Models · ContextEvaluation Engineering, Trace Replay, and Run Comparison
Agent behavior changes when the model, prompt, tools, harness, context, environment, repository, or evaluator changes. A handful of successful demos cannot show whether a configuration is reliable, whether a new version is better, or whethe
Plan · Execute · Verify · Learn · Agent Runtime · Models · ContextAgent and Loop Engineering Patterns
An agent can reason, act, observe, and continue, but production work needs more than an open ended loop. The system must decide which tasks deserve agents, which agent configuration is eligible, when to use one agent or several, what happen
Plan · Execute · Verify · Learn · Agent Runtime · Models · ContextEvaluation Science and Controlled Experimentation
Add experimental rigor to repeatable agent evaluation and promotion decisions.
Verify · Learn · Agent Runtime · Models · ContextCapability Learning, Optimization, and Regression Control
Turn failures, corrections, and successful strategies into controlled capability improvements.
Learn · Define · Verify · Agent Runtime · Models · ContextKnowledge, Context, and Retrieval Pipeline Specification
Specify the governed path from registered source to immutable, attributable context package.
Plan · Execute · Verify · Learn · Agent Runtime · Models · ContextMulti-Agent Topologies and Collaboration Contracts
Select and govern multiple agents only when specialization, parallelism, or independent assurance creates measurable value.
Plan · Execute · Verify · Learn · Agent Runtime · Models · ContextAgentic Architecture Patterns and Autonomy Selection
Select the minimum sufficient architecture for the problem and risk rather than maximizing how agentic the design appears.
Intent · Plan · Execute · Verify · Learn · Agent Runtime · Models · ContextAutonomous Workflows
4 documentsRepository Onboarding and Codebase Intelligence
Establish the evidence required before an autonomous workflow may operate on a repository.
Intent · Plan · Execute · Orchestration · WorkflowAutonomous Engineering Workflow Catalog
Turn “use agents for engineering” into a portfolio of explicit, governable workflow products.
Intent · Plan · Execute · Verify · Deliver · Learn · Orchestration · WorkflowChange Workflows — Features, Defects, Tests, and Modernization
Define distinct operating patterns for the most common code changing workflows.
Intent · Plan · Execute · Verify · Deliver · Orchestration · WorkflowOperational Workflows — Security, Incidents, Production, and Knowledge
Define factory workflows that begin from operational or knowledge signals rather than a planned feature.
Execute · Verify · Deliver · Learn · Orchestration · WorkflowVerification & Delivery
3 documentsSoftware Testing Strategy for Agentic Change
Define a complete testing system for changes produced or maintained by agents.
Plan · Execute · Verify · Verification · DeliveryCI/CD, Artifacts, Migrations, and API Compatibility
Connect governed agent work to reproducible builds, immutable artifacts, compatibility controls, and delivery systems.
Verify · Deliver · Verification · DeliveryProgressive Delivery, Production Verification, and Rollback
Complete the factory path from release eligibility to observed customer outcome.
Deliver · Learn · Verification · DeliveryFactory Platform
8 documentsDeveloper Portal, Service Catalog, and Golden Paths
Treat the software factory as an internal product that developers and agents can discover, understand, and use safely.
Intent · Plan · Execute · Verify · Deliver · Learn · Platform · OperationsScheduling, Capacity, Cost, and Fairness
Govern scarce model, compute, environment, tool, and human review capacity.
Execute · Platform · OperationsResilience, Disaster Recovery, and Factory SRE
Keep the factory safe and explainable when its own infrastructure fails.
Execute · Verify · Deliver · Learn · Platform · OperationsHuman-Agent Control Surfaces and Operator Experience
Design interfaces that let people understand, direct, interrupt, and judge autonomous work without reading raw logs.
Intent · Plan · Execute · Verify · Deliver · Learn · Platform · OperationsWorkflow and Event Contracts, Schema Evolution, and Factory Storage
Define the durable contracts that connect triggers, workflow state, asynchronous events, artifacts, evidence, and retrieval stores.
Plan · Execute · Verify · Learn · Platform · OperationsObservability Semantics, Cost Attribution, and Forensics
Make runs comparable and incidents reconstructable across agents, tools, services, queues, environments, CI, deployment, and outcomes.
Execute · Verify · Deliver · Learn · Platform · OperationsEnterprise Operations, Reliability, and FinOps Reference
Consolidate the operating contract for admission, capacity, budgets, reliability, continuity, incidents, and cost per accepted outcome.
Execute · Verify · Deliver · Learn · Platform · OperationsControl Tower Monitoring, Detection, and Response
Connect inventory, authority, health, quality, safety, cost, drift, incidents, response, and verified closure in one operating view.
Execute · Verify · Deliver · Learn · Platform · OperationsQuality Engineering
5 documentsQuality and Evidence Architecture
Define how the factory proves that an exact candidate satisfies exact requirements under known conditions.
Verify · Deliver · Learn · Quality · EvidenceRelease, Production Feedback, and Factory SRE
A review ready pull request is not customer value. Merge, deployment, runtime health, rollback, and outcome confirmation remain separate claims. A factory that stops at code generation can accelerate inventory while ignoring whether the cha
Verify · Deliver · Learn · Quality · EvidenceContinuous Quality Contracts, Proof Packages, and Certificates
This chapter extends Quality and Evidence Architecture. That chapter defines evidence semantics. This chapter explains how to compile requirements into a continuous quality control system spanning the complete software lifecycle.
Verify · Deliver · Learn · Quality · EvidenceQuality Contract and Certificate Technical Specification
“Produce a quality proof package” is architecture doctrine, not yet an implementable contract. A real system needs stable subjects, schemas, lifecycle states, evidence sufficiency rules, deterministic decisions, signatures, invalidation, re
Verify · Deliver · Learn · Quality · EvidenceProduction Feedback, Reproduction, Automated Review, and Merge
User feedback is valuable and incomplete. A report may describe an obsolete version, duplicate another symptom, omit the operating conditions, or attribute the failure to the wrong component. Creating engineering issues directly from raw fe
Verify · Deliver · Learn · Quality · EvidenceSecurity & Governance
7 documentsGovernance, Policy, and Risk-Proportional Approval
An autonomous agent can possess the technical capability to modify a repository, call an API, or trigger a delivery system without possessing the organizational authority to do so. That distinction is the beginning of governance.
Intent · Execute · Verify · Deliver · Security · GovernanceSecurity and Identity Architecture
An AI Software Factory connects human intent to code, credentials, repositories, tools, and delivery systems. A confused identity boundary can let a browser pretend to be an agent, a service inherit human authority, or a repository credenti
Intent · Execute · Verify · Deliver · Security · GovernanceSoftware Supply Chain Security, Provenance, and Attestation
Passing tests does not establish what was built, from which source, by which identity, with which dependencies, or whether the artifact later changed. An autonomous factory expands the supply chain: models, prompts, tools, MCP servers, runn
Intent · Execute · Verify · Deliver · Security · GovernanceAgentic Threat Model and Adversarial Defense
Threat model systems in which models interpret untrusted content and can take multi step action through tools.
Plan · Execute · Verify · Deliver · Security · GovernanceWorkload Identity, Secrets, Privacy, and Compliance
Define who or what acts, which authority it receives, how credentials are delivered, and how information obligations persist across the factory.
Intent · Execute · Verify · Deliver · Learn · Security · GovernanceAgentic Governance Control Framework
Convert governance principles into ten testable control families.
Intent · Plan · Execute · Verify · Deliver · Learn · Security · GovernanceAuthority, Autonomy, and Emergency Control
Define how authority is delegated, exercised, reduced, revoked, contained, and restored.
Intent · Execute · Verify · Deliver · Learn · Security · GovernanceCase Studies
3 documentsMission Control Implementation Maturity and Evidence Map
Historical assessment: this map captures the repository states listed below. For the verification architecture at merged commit ff0524e , continue with the Verification First Software Factory case study. For the current checkout at d902fae
Execute · Verify · Control Plane · EvidenceVerification-First Software Factory — Mission Control Case Study
An agent can generate code, run a test, and confidently report success while still misunderstanding the request, exceeding its authority, weakening the test system, or validating a different artifact from the one placed in a pull request. T
Execute · Verify · Control Plane · EvidenceMission Control Capability, Workflow, and Admission Map
Evidence boundary: this assessment uses the tracked Mission Control files at checked out commit d902fae . The checkout was detached from a branch, and its local origin/main ref pointed to 4700573 . Untracked review artifacts were excluded.
Execute · Verify · Control Plane · EvidenceLabs
13 documentsGoverned Issue to Validated Pull Request
Prove that Mission Control can govern one bounded software change from human intent to a review ready pull request. The learner must operate, trace, explain, validate, and recover the workflow. Autonomous deployment is outside scope.
Execute · Verify · Learn · PracticeCapstone Architecture and Executive Defense
The capstone proves personal mastery of the supported Mission Control V1 path, not aspirational factory autonomy. It passes only when you can truthfully say:
Execute · Verify · Learn · PracticeCapability Certification and Revocation Lab
Prove that an agent capability can be packaged, evaluated, certified for a bounded scope, resolved into an execution contract, and revoked without losing historical explainability.
Define · Verify · Learn · PracticeRepository Onboarding and Readiness Lab
Demonstrate that repository registration, discovery, owner validation, indexing, and workflow admission are separate, evidence backed steps.
Intent · Plan · PracticeAgentic Security Attack and Containment Lab
Prove that untrusted repository content and tool output cannot widen authority, expose secrets, poison durable memory, or corrupt evidence.
Execute · Verify · Learn · PracticeProgressive Delivery and Rollback Lab
Demonstrate the complete path from immutable release candidate through canary, production verification, failure detection, rollback or containment, and outcome accounting.
Deliver · Learn · PracticeIncident Remediation and Postmortem Lab
Use agents to accelerate evidence collection and correction while preserving incident authority, forensic integrity, and clear separation between observation, hypothesis, containment, and root cause.
Execute · Verify · Deliver · Learn · PracticeContinual Improvement Promotion Lab
Turn recurring production like feedback into an evaluated, human approved capability improvement without allowing the system to mutate active behavior directly.
Learn · Define · Verify · PracticeFactory Disaster Recovery Lab
Prove that the factory can restore authoritative state, prevent split brain execution, reconcile uncertain external effects, and return safely to service after a major failure.
Execute · Verify · Learn · PracticeAuthority, Containment, and Decision Replay Lab
Prove that a factory can stop unsafe work through an independent control path, reconstruct delegated authority without hidden model reasoning, reconcile partial effects, and restore only the minimum qualified authority.
Execute · Verify · Learn · PracticeOrchestration Failure, Recovery, and Cost Lab
Prove bounded convergence and safe recovery under provider degradation, unknown side effects, budget pressure, duplicate delivery, partial multi step results, and human escalation. Attribute the whole cost to the accepted or failed outcome.
Execute · Verify · Learn · PracticeKnowledge Poisoning, Revocation, and Retrieval Lab
Demonstrate permission filtering before ranking, stale and contradictory source handling, poisoning containment, reverse lineage impact analysis, revocation propagation, deletion, clean reindexing, and reproducible context packages.
Plan · Execute · Verify · Learn · PracticeExternal Capability Intake and Recertification Lab
Qualify a synthetic external tool through ownership, provenance, data use, security, service, side effect, cost, incident, exit, certification, publication, drift, revocation, and recertification controls.
Define · Execute · Verify · Learn · PracticeInterview Practice
1 documentResearch Journal
1 documentReference
2 documentsAI Software Factory Mastery Curriculum
The curriculum moves from the purpose of the factory to its operating model, domain, runtime, assurance systems, implementation, and leadership use. The sequence matters. Runtime mechanisms make sense only after the learner understands the
Chapter Writing Standard
This guide should read like a graduate level engineering textbook or a serious technical leadership book. Its chapters should build durable mental models, show why systems take their current shape, and develop the reader's ability to reason