engineering-standards

Use only when the request concerns standards, responsibility, complexity, smells, naming, refactoring catalogs, or task-plan rules; not routine work.

SKILL.md

GitHub

Skill: engineering-standards

Purpose

Apply the shared engineering standards library: manifesto, constitution, design principles, responsibility model, complexity budget, smells, naming, testing, errors, observability, API/domain/state/config/data/dependency philosophy, and evolution rules.

Use this skill when the task is broad, architectural, quality-focused, standards-focused, or when a change risks growing into a dumping ground.

Required references

Read only the references needed for the task:

  • ../../engineering/README.md — index and selection rules for the standards library.
  • ../../engineering/00_manifesto.md — stable engineering intent.
  • ../../engineering/01_constitution.md — non-negotiable engineering law.
  • ../../engineering/05_design_principles.md — SOLID, KISS, DRY, YAGNI, Clean Architecture, DDD, TDD as behavior.
  • ../../engineering/06_responsibility_model.md — single responsibility at function/class/file/package/service/layer/context levels.
  • ../../engineering/07_complexity_budget.md — size and complexity thresholds.
  • ../../engineering/08_engineering_smells.md — code-level smells.
  • ../../engineering/09_architectural_smells.md — boundary and architecture smells.
  • ../../engineering/10_antipattern_catalog.md — recurring bad design patterns.
  • ../../engineering/11_refactoring_catalog.md — safe refactoring tools.
  • ../../engineering/12_naming_bible.md — naming rules.
  • ../../engineering/24_task_todo_style.md — structured TODO/task plans.
  • ../../engineering/25_api_philosophy.md — API and CLI contract design.
  • ../../engineering/26_domain_modeling.md — domain boundaries, invariants, commands, events, and policies.
  • ../../engineering/27_state_machine_philosophy.md — explicit states and valid transitions.
  • ../../engineering/28_concurrency_philosophy.md — async, workers, locks, queues, retries, timeouts, and backpressure.
  • ../../engineering/29_configuration_philosophy.md — configuration, secrets, flags, defaults, and lifecycle.
  • ../../engineering/30_dependency_philosophy.md — dependency admission, wrapping, risk, and removal.
  • ../../engineering/31_data_philosophy.md — source of truth, schema changes, migrations, retention, caching, and consistency.
  • ../../engineering/32_ui_architecture_philosophy.md — UI state, component boundaries, accessibility, and frontend architecture.
  • ../../engineering/33_ai_engineering_philosophy.md — evidence-bound AI engineering and agent patch quality.
  • ../../engineering/34_evolution_decision_tree.md — decision trees for adding structure.
  • ../../engineering/35_evidence_contract.md — evidence requirements and uncertainty reporting.
  • ../../engineering/36_task_lifecycle_gates.md — proportional task gates.
  • ../../engineering/37_review_regression_gates.md — review and regression gates.
  • ../../engineering/38_library_drift_audit.md — library drift audit contract.

Operating rules

  • Do not apply principles as slogans.
  • Translate principles into concrete design choices.
  • Do not force Clean Architecture, DDD, CQRS, event sourcing, or similar styles unless the repository already uses them or the user asks.
  • Prefer the smallest design that preserves correctness and future maintainability.
  • Treat large files and mixed responsibilities as risks, not style issues.
  • A human-authored 10k-line file is unacceptable except generated/vendor/lock/migration-like exceptions or explicit legacy containment.

Output

When this skill materially affects the answer, name the principle or smell and the concrete action taken.