Conciseness

The body is lean and assumes competence — the Quick Reference, the modes table, and short focused sections each earn their tokens — but the instruction "Always use mode: 'structured' instead of mode: 'tables'" is repeated three times (table row "Default choice — use this", the bold paragraph, and the Anti-Patterns entry), and Chart Hints/Brand Mode content overlaps with the modes table. That repetition puts it at level 4 (efficient, minor instances that could be trimmed) rather than level 5 where every token earns its place.

Actionability

The Quick Reference gives copy-paste-ready JSON invocations for the main modes and Named Schemas supplies a concrete example, so most guidance is executable. However, some sections only describe behavior without a usable example — `max_tokens_out` has no sample value or effect illustration, and there is no example of the output shape — which are minor gaps consistent with level 4 rather than the fully-covered common cases of level 5.

Workflow Clarity

This is a simple single-action skill (invoke the extract tool with parameters), and the single action is unambiguous: the modes table, the bolded default ("Default choice — use this"), the Anti-Patterns, and the "When NOT to use" section together give an explicit decision path with no risky multi-step sequence requiring validation checkpoints. The simple-skill exception applies, so it reaches 5; level 4 would be appropriate only if mode selection were ambiguous.

Progressive Disclosure

The body is well organized into short overview-level sections with nothing that belongs in a separate file, and the bundle has no reference files to split out — good structure overall. The deduction is for the "See Also" links pointing at `../wigolo-fetch/SKILL.md` and `../wigolo-agent/SKILL.md`, which do not resolve from the skill's location in this evaluation context, leaving navigation partially broken — a minor organization gap placing it at level 4 rather than the clean, easily navigable level 5.

Specificity

The description enumerates concrete extraction targets — "tables, definition lists, key-value pairs, JSON-LD, microdata, chart hints (SVG titles / aria-labels / figcaptions), brand assets, and metadata" — giving comprehensive, specific coverage of what the skill does rather than vague language. It matches the level-5 anchor (multiple specific concrete actions, comprehensive coverage) and exceeds level 4, which allows minor coverage gaps.

Completeness

Both questions are answered explicitly: the "what" is the enumerated structured-extraction capability, and the "when" is a concrete "Use when the user wants structured data... or says 'extract the table'..." clause. It even adds boundary guidance ("For autonomous navigation across many pages, use wigolo's `agent` tool instead"), which goes beyond the level-5 example; level 4 would mean the when-clause was present but less explicit.

Trigger Term Quality

It includes both topical keywords ("structured data, pricing tables, feature comparisons") and verbatim natural user utterances — "extract the table", "get structured data", "pull the pricing", "extract as JSON". This is comprehensive natural-term coverage with synonym variation, matching the level-5 anchor; level 4 would require noticeable missing everyday phrasings, and none are evident.

Distinctiveness Conflict Risk

The niche is clear — structured data extraction from webpages — and the description actively disambiguates against the sibling skills ("For autonomous navigation across many pages, use wigolo's `agent` tool instead"), minimizing wrong-skill triggering. It matches the level-5 anchor (clear niche with distinct triggers); level 4 would imply unresolved overlap with closely related skills such as the fetch/markdown tool, which the boundary hint partly addresses.