engineering-everything

Use when choosing technical architecture, project shape, stack, data source, boundaries, deployment model, AI/Agent architecture, or scaffold strategy.

Use when designing or implementing RPA, OCR, CV, LLM production pipelines, browser automation, data governance, analytics reporting, or POC/spike automation.

Use when implementing, fixing bugs, migrating code, wiring APIs, running validation commands, or closing a build/test loop with minimal scoped changes.

Use when a task needs engineering-route judgment before choosing product definition, project inheritance, architecture, execution, build, refactor, review, release, organization, automation, or skill-evolution work.

Use when splitting a feature, writing an implementation plan, choosing worktree strategy, defining vertical slices, validation gates, or review roles before coding.

Use when engineering an organization, company mechanism, SOP, onboarding, interview process, non-software project, WBS, estimate, workspace, knowledge base, table, or document-led operating system.

Use when clarifying a product idea, PRD, requirements, PSPS, user scenario, scope, non-goals, or success criteria before architecture or implementation.

Use when taking over an existing repository, auditing an old project, deciding whether to refactor, or planning how to continue work from current code and documents.

Use when refactoring, cleaning up, modularizing, renaming, reducing technical debt, or preserving behavior while improving structure.

Use when reviewing code, PRs, merge readiness, CI failures, release risk, rollback plans, or deciding whether a change is safe to ship.

Use when evolving a skill, capturing lessons, promoting patterns, changing routes, updating references, installing skills, validating skill packages, or publishing skill releases.

Use when starting, resuming, or switching an Engineering Everything session before selecting the kernel router or a direct engineering sub-skill.