Conciseness

The body is mostly lean and command-driven with no explanation of concepts Claude already knows, but several sections are tightening candidates: the "Industry-Specific E-E-A-T Thresholds" table (8 industries) and "Anti-Patterns Rejected" substantially duplicate content already in references/aeo_eeat_canon.md, the "Trigger Phrases" section repeats the frontmatter description verbatim in spirit, and the porting-history footer ("2,464 LOC across 9 modules …") adds little operational value. This fits anchor 3 (mostly efficient but could be tightened) better than anchor 2, since the duplicated material is a minority of the file and everything else earns its tokens.

Actionability

Every tool has a copy-paste-ready invocation with real flags and example values: "python3 scripts/aeo_audit.py --url https://example.com/blog/post", "aeo_optimizer.py --input post.md --mode balanced --output post-aeo.md", and a fully parameterized citation_tracker add/report pair; the three scripts all exist in scripts/. Concrete output samples (the markdown report template) and a worked configuration example (--industry healthcare) cover the common cases — anchor 5.

Workflow Clarity

The workflow is a clearly numbered 0–4 sequence, each step paired with a concrete command and its expected result ("→ markdown report with composite score + 4-dimension breakdown"), and step 0 is an explicit gating checkpoint ("a blocked GPTBot/PerplexityBot/ClaudeBot/Google-Extended is the first fix, always"). It falls short of anchor 5 because there are no feedback loops (e.g., re-audit after applying fixes, or verify the optimizer's output before publishing); the operations are non-destructive (optimizer writes to a new output file), so the destructive-operation cap at 3 does not apply.

Progressive Disclosure

The References section clearly signals all five one-level-deep reference files with one-line descriptions of each, and all five exist in references/ (verified), as do the three scripts named in Core Capabilities. The gap keeping it at anchor 4 rather than 5 is that the industry-thresholds table and anti-patterns list are inlined in SKILL.md despite being covered by references/aeo_eeat_canon.md, and the Trigger Phrases section duplicates the frontmatter rather than pointing into the body's structure.

Specificity

The description lists multiple concrete actions — "optimize content to be cited by AI language models (ChatGPT, Perplexity, Claude, Gemini, Mistral) as authoritative sources", "auditing existing content for E-E-A-T signals", "tracking which pages get cited by which LLMs", "Output is a markdown audit report (default) or JSON for pipeline integration" — with concrete deliverables and tool constraints ("Stdlib-only Python tools"). This matches the top anchor (multiple specific concrete actions, comprehensive coverage) and exceeds anchor 4, which allows coverage gaps.

Completeness

Both questions are explicitly answered: what ("optimize content to be cited by AI language models … as authoritative sources", distinct from SEO, with stated output formats) and when ("Use when planning content for AI-first search audiences, auditing existing content for E-E-A-T signals, tracking which pages get cited by which LLMs, or building a citation-friendly content strategy"), with concrete trigger phrases. This is the anchor-5 pattern; anchor 4 would require the 'when' to be less explicit.

Trigger Term Quality

It enumerates seven natural trigger phrases users would actually say: "AEO audit", "optimize for ChatGPT", "get cited by Perplexity", "LLM citation strategy", "answer engine optimization", "content for AI search", "E-E-A-T audit", plus additional use-when scenarios. Coverage includes the canonical term, its abbreviation, platform-specific and signal-specific phrasings — comprehensive with synonyms, matching anchor 5 rather than anchor 4 ("a few natural terms missing").

Distinctiveness Conflict Risk

It carves out a clear niche and explicitly disambiguates from its nearest overlapping skill: "Distinct from SEO — AEO optimizes for citation in LLM-generated responses, not search rankings." The trigger terms (AEO audit, E-E-A-T audit, LLM citation strategy) are specific to this domain, minimizing wrong-skill triggering — anchor 5 rather than anchor 4 (minor overlap risk).