Conciseness

The Check and Fix sections are dense and instruction-first, but the motivational intro ("Images are typically 50%+ of page weight... These savings compound across every visitor"), the Explain section's tutorial on WebP/AVIF compression and browser support that Claude already knows, and statistics stated twice ("25-35%" and "40-50%" appear in both Quick Reference and Explain) are unnecessary padding that could be merged. This sits between the 2 anchor (several padded sections) and 4 anchor (minor trim only), so 3 fits best.

Actionability

Guidance is concrete and executable in prose: scan targets are enumerated ("Any <img src> pointing to .jpg, .jpeg, or .png", CSS background-image, missing <source> entries), fixes specify quality settings ("WebP at 80% quality using Squoosh or Sharp", "AVIF at 60% quality"), the exact <picture>/<source> structure, the JPEG fallback, and a "@supports (background-image: url('.webp'))" pattern. As an instruction-only skill this is actionable without code, but a copy-paste <picture> markup example or a concrete Sharp command would lift it to 5.

Workflow Clarity

The Check → Fix → Explain → Code Review sequence is clear and the steps are numbered with reporting requirements ("file path and line number"), which would rate 4. However, the Fix step is a batch operation across all images and markup, and the workflow contains no explicit validation checkpoint for Claude's own changes — the DevTools confirmation appears only as something to "describe" to the user in the Code Review section — so per the batch-operation cap, workflow clarity cannot exceed 3.

Progressive Disclosure

The ~30-line body is organized into scannable sections (Quick Reference, Check, Fix, Explain, Code Review) and defers detail via a clearly signaled one-level-deep pointer: "For full implementation details, code examples, and framework-specific guidance, see references/rule.md" — a real 189-line file with no nested references. This matches the clear-overview-with-well-signaled-references anchor.

Specificity

The description lists several concrete review actions and targets — "reviewing image assets, markup, and CDN or build transforms" and "Check encoded size, rendered size, loading strategy, and above-the-fold impact together" — which is specific and multi-faceted. It falls short of a 5 because it stays at the review level and never names the remediation actions the skill actually performs (convert, rewrap markup), leaving minor coverage gaps.

Completeness

Both what ("reviewing image assets, markup, and CDN or build transforms... Check encoded size, rendered size, loading strategy, and above-the-fold impact") and when ("Use when reviewing...") are present and explicit. Not a 5 because the 'when' clause is anchored to the rule title ("related to Use modern image formats (WebP, AVIF)") rather than concrete user-facing trigger phrases like 'user mentions WebP/AVIF or image formats'.

Trigger Term Quality

Natural terms a user would say are present: "image assets", "markup", "CDN or build transforms", "WebP, AVIF", "above-the-fold". It misses a few common variations and file extensions users actually type (.webp, .avif, .jpg/.png, "image optimization", "page weight"), which keeps it below the comprehensive anchor at 5 but clearly above the 3 anchor's thin keyword coverage.

Distinctiveness Conflict Risk

The WebP/AVIF format-review niche with CDN/build-transform and above-the-fold framing is mostly distinct from sibling image skills. Minor overlap risk remains with closely related image skills (lazy loading, responsive sizing, image CDN configuration) that also trigger on "image assets" and "loading strategy".