Conciseness

The 'Shared project skill-context' preamble repeats the same instruction three ways (the ordered read list, the repeated 'Read .ai-factory/skill-context/aif-best-practices/SKILL.md — MANDATORY', the precedence bullets, and the enforcement paragraph), and the core sections restate basic programming knowledge Claude already has ("Boolean variables: is, has, can prefix", "Max 20-30 lines per function", single responsibility, the AAA pattern). With both padded meta-instructions and textbook restatements, this matches anchor 2 rather than anchor 3 — the format is tight, but the content does not assume Claude's competence.

Actionability

The body provides executable TypeScript/PHP code (calculateDiscount with a guard clause, a UserNotFoundError class, an AAA-pattern test), ✅/❌ contrast tables, a coverage-priority chart, and concrete checklists. Minor abstract rules remain ("Avoid circular dependencies", "Use error boundaries at system edges"), keeping it at anchor 4 rather than 5.

Workflow Clarity

Mode routing is unambiguous (Quick Reference maps each argument to a section), the skill-context reads are an ordered sequence, and there is an explicit post-output verification step ("After generating any output artifact, verify it against all skill-context rules"). As an advisory, read-only skill with no destructive or batch operations, no validation cap applies; small gaps (e.g., the duplicated context-read instruction) hold it at anchor 4.

Progressive Disclosure

No bundle files (references/, scripts/, assets/) exist, and the single-file body is well organized: clear section headers, a Quick Reference navigation, and a summary table, with external file references (.ai-factory/skill-context/..., .ai-factory/ARCHITECTURE.md) explicitly signaled and one level deep. Anchor 4 fits — good structure with minor gaps (the ~290-line body could split per-topic content into reference files, and the skill-context instructions are duplicated) rather than the clean split-and-navigate pattern of anchor 5.

Specificity

The description lists five specific coverage areas ("Covers naming, structure, error handling, testing, and code review standards") beyond naming the domain, though they are framed as topics rather than concrete actions and "Code quality guidelines and best practices" is generic. This matches anchor 4 (several specific items, minor gaps) — above anchor 3 (only 1-2 concrete actions) but below anchor 5 (comprehensive concrete actions).

Completeness

Both parts are explicit: the what ("Code quality guidelines and best practices... Covers naming, structure, error handling, testing, and code review standards") and the when with concrete trigger phrases ("Use when writing code, reviewing, refactoring, or asking..."). This directly matches anchor 5; the explicit 'Use when...' clause means no cap applies.

Trigger Term Quality

"Use when writing code, reviewing, refactoring, or asking 'how should I name this', 'best practice for', 'clean code'" includes natural phrases users would actually say. It falls short of anchor 5 because common synonyms like "code style", "conventions", or "lint" are missing.

Distinctiveness Conflict Risk

The primary trigger "Use when writing code" is broad enough to overlap with code-review, security, and language-style skills, though "clean code" and "best practice for" carve out a partial niche. This fits anchor 3 (somewhat specific but could still overlap with similar skills) rather than anchor 4, because the broad writing/reviewing triggers create real overlap risk.