software-architecture-engineering-uap

Software Architecture Engineering UAP

Core Rule

Treat architecture as traceable engineering: requirements -> context -> schematic style choices -> architecture views -> NFR tactics -> evaluation -> implementation work. Do not recommend structure, technology, or patterns without naming the forces, trade-offs, evidence, and validation path.

First Pass

1. Classify the task: greenfield design, architecture review, modernization, decomposition, clean-code review/refactoring, Rust library/crate architecture, microservices pattern design, API/service design, data architecture, IAM/auth design, security-pattern design, cost-aware decision, deployment/operation planning, architecture-as-code governance, NFR design, evaluation, ADR/RFC, or implementation planning.
2. Identify scope and risk: users, stakeholders, business goals, current system evidence, constraints, critical workflows, quality attributes, team/deployment context, and reversibility.
3. Tailor the Unified Architecture Process: lightweight for local low-risk changes; full A1-A6 for high-risk, distributed, regulated, long-lived, or irreversible decisions.
4. Decide the visual artifacts needed: boundary/context, use case, component, class/data, activity, sequence, state, deployment, development/operation, style schematic, NFR conformance map, policy enforcement map, or fitness-function map.
5. Define verification before recommending: repository inspection, docs review, scenario walk-through, prototype, benchmark, threat model, operational review, or implementation test plan.

Required Reads By Task

• Intake, scoping, or deciding how much process to use: tasks/intake-and-tailoring.md.
• Ambiguous requirements, stakeholder conflict, or missing acceptance criteria: tasks/requirements-refinement.md.
• External actors, integrations, data flows, system boundaries, or operational environment: tasks/system-context-analysis.md and references/context-and-views.md.
• Architecture style choice, schematic structure, decomposition, or modernization target: tasks/schematic-architecture-design.md and references/architecture-style-catalog.md.
• Microservices, modular monolith vs microservices, service decomposition, saga compensation/retry/idempotency, outbox relay/deduplication, CQRS projection/rebuild/lag, API composition, API gateway, BFF, service discovery, circuit breaker, event schema/order/replay/DLQ, contract testing, service chassis, service mesh, sidecar, strangler migration, anti-corruption layer, or production-ready service design: tasks/microservices-pattern-architecture-design.md and references/microservices-pattern-language.md.
• Design principles, modularity, coupling/cohesion, KISS, SOLID, DRY, YAGNI, SoC, package boundaries, or code-level architecture heuristics: references/design-principles-and-modularity.md.
• Clean code, code smells, refactoring, code review, craftsmanship, naming/functions/classes/comments/error handling/tests, behavior-preserving cleanup, technical debt reduction, or implementation hygiene: tasks/clean-code-refactoring-and-review.md, references/clean-code-craftsmanship.md, and templates/clean-code-review-checklist.md.
• Rust library, crate, SDK, parser, reusable package, public API design, Cargo features, MSRV, SemVer, unsafe, FFI, no_std, macro crate, async crate, or senior QA/testing plan for a Rust library: tasks/rust-library-design-review-and-testing.md, references/rust-library-architecture.md, templates/rust-library-contract-matrix.md, and templates/rust-library-release-flow.mmd.
• Functional, information, behavior, deployment, development, or operation design: tasks/architecture-view-design.md, references/context-and-views.md, references/development-and-operation-views.md, and references/diagram-visualization-guide.md.
• Identity, IAM, authentication, authorization, sessions, tokens, MFA, passwordless, account lifecycle, admin users, audit logs, or auth API design: tasks/identity-access-design.md and references/iam-auth-architecture.md.
• Security patterns, access-control models, policy engines, reference monitor, RBAC/ABAC/PBAC/ACL/capability, security logger/auditor, secure middleware/network patterns, misuse cases, or threat-driven design: tasks/security-pattern-architecture-design.md and references/security-architecture-patterns.md.
• Cost, estimation, schedule/effort risk, operational cost, total ownership cost, performance capacity, resource-usage modeling, PERT/WBS/function points/COCOMO-style reasoning, or queueing models: tasks/cost-aware-architecture-decision.md and references/cost-estimation-and-performance-models.md.
• Architecture-as-code, ADL-lite constraints, dependency rules, layer/database constraints, fitness functions, repository/package alignment, or architecture drift: tasks/architecture-as-code-governance.md and references/architecture-as-code-and-fitness-functions.md.
• Performance, security, availability, modifiability, observability, compliance, or other NFR work: tasks/nfr-and-conformance-design.md and references/nfr-tactics-and-conformance.md.
• Architecture evaluation, review gate, design risk assessment, or validation plan: tasks/architecture-evaluation.md and references/architecture-evaluation-methods.md.
• Existing system review, refactor, migration, or modernization: tasks/architecture-review-and-modernization.md.
• ADR/RFC writing, implementation breakdown, rollout, tests, or repository mapping: tasks/implementation-planning.md and references/engineering-translation.md.
• When a prompt primarily belongs to another skill but code-level maintainability is material, use the cross-skill map in references/clean-code-craftsmanship.md; do not load clean-code material for pure strategy, research, or architecture-only prompts without code/refactoring/testability impact.
• Quick UAP activity-step-task reminders or SRS intake extraction: references/uap-appendix-artifact-playbooks.md.
• Final architecture quality check: references/review-checklist.md.