The body is mostly lean with well-organized sections, but the opening sentence ('An incorrect noindex tag can completely remove important pages from search results, while proper usage helps manage crawl budget...') explains a concept Claude already knows. Not a 1 because the rest is tight; not a 3 because that explanatory preamble does not earn its tokens.

The 'Fix' section gives a concrete action and the 'Code Review' section tells the model to flag exact routes, but the 'Check' step is vague ('Verify that the noindex directive is only applied to pages...') and the meta-tag example is HTML-entity-escaped ('<meta name=...>'), which is not copy-paste ready. Not a 1 because some concrete guidance exists; not a 3 because the executable detail is incomplete and one example is malformed.

The Check/Fix/Explain/Code Review sections provide a loose sequence, but there are no validation checkpoints or verify-then-retry feedback loops in the body for what is effectively a batch audit. Per the rubric, missing validation for batch operations caps this at 2. Not a 1 because a sequence is present; not a 3 because validation gates are absent from the body.

The body is a concise overview that clearly signals one-level-deep detail ('For full implementation details, code examples, and framework-specific guidance, see references/rule.md'), and that reference file exists and is organized. Not a 2 because navigation is explicit and the split is appropriate; the structure matches the top anchor.

Names the SEO domain and concrete actions ('auditing metadata, crawlability, structured data, or indexability', 'Verify the rendered HTML and HTTP response'), but the capabilities are broad rather than a comprehensive list of specific actions. It is not a 1 because real actions are named, and not a 3 because the action set is partial and the awkward 'Audit all noindex pages' insertion muddies what is actually being done.

It has an explicit 'Use when...' trigger clause and states what it does ('auditing metadata, crawlability, structured data, or indexability' plus verifying rendered HTML/HTTP responses), answering both what and when. It is not a 2 because both halves are explicit rather than implied.

It includes relevant trigger terms like 'metadata, crawlability, structured data, or indexability' that a user might say, but misses common variations such as 'noindex', 'robots meta tag', 'blocked from Google', or 'de-indexed'. Not a 1 because natural SEO keywords are present; not a 3 because coverage of the terms users would actually say is incomplete.

The 'noindex pages' focus gives it a niche, but the description leans on broad terms ('metadata, crawlability, structured data, or indexability') that overlap with other SEO skills. Not a 1 because it is anchored to a specific rule; not a 3 because the wording is generic enough to collide with sibling SEO audit skills.