At roughly 1000 lines the body restates the same rules across the Quick Start checklists, the per-section Rules blocks, the Core Principles, and the Anti-Patterns table, so it could be materially tightened. Not score 1 because it mostly avoids explaining basic concepts Claude already knows, but not score 3 due to the redundancy and overall length.

Provides extensive copy-paste-ready executable code in TypeScript, Python, and Go plus concrete commands (prisma migrate, alembic upgrade, curl health checks) and decision tables. Not score 2 because examples are complete and executable rather than pseudocode.

The MANDATORY WORKFLOW sequences Steps 0–5 with an explicit Step 4 'Test & Verify' validation block (build, smoke, integration, real-time checks) and the directive 'If any check fails, fix the issue before proceeding.' Not score 2 because explicit validation checkpoints and a feedback loop are present.

Eight real one-level-deep reference files exist and are surfaced via Quick Navigation and Reference Documents tables, but the SKILL.md body itself is a ~1000-line inline reference rather than a concise overview — content that arguably belongs in references is inline. Not score 1 because references are genuine and well-signaled, but not score 3 because the main file is far from a lean overview.

Lists multiple concrete actions such as "scaffolding backend service", "implementing auth flows", "handling file uploads", and "adding real-time features (SSE/WebSocket)"; uses third-person gerunds with no first/second-person voice to penalize. Not score 2 because the action list is comprehensive rather than partial.

Explicitly answers what ("Full-stack backend architecture and frontend-backend integration guide") and when via a literal "TRIGGER when:" block plus a "DO NOT TRIGGER when:" block, satisfying the explicit-trigger requirement. Not capped at 2 because explicit trigger guidance is present.

Covers natural phrasing a user would actually say — "building a full-stack app", "building todo app", "building CRUD app", "Express + React", "Next.js API" — across both task and stack terms. Not score 2 because common variations are well represented rather than partially missing.

Broad pure-backend triggers ("Node.js backend", "Python backend", "Go backend", "designing service layers", "implementing error handling") would overlap with dedicated backend skills, and the DO NOT TRIGGER clause excludes only frontend/CSS/schema, not pure-backend work. Not score 3 because conflict with backend-only skills remains likely.