Conciseness

The body is ~60 words with no padding and no explanation of concepts Claude already knows — every token is a directive ('Read `.claude/skills/no-ai-slop/SKILL.md` and execute it exactly as written'). Brevity is maximal, though it achieves this by containing almost no content.

Actionability

The only concrete guidance is which files to read ('Read `.claude/skills/no-ai-slop/SKILL.md`', 'follow `.agents/rules/cog.md`', 'invoke `.agents/agents/<name>.md` via `invoke_subagent`'); the actual task steps — how to edit drafts or which AI-slop patterns to detect — are entirely delegated to files outside the bundle. This matches anchor 2 (high-level pointers, missing the specific steps to execute), not 3, because no task-relevant executable guidance exists in the body at all.

Workflow Clarity

A rough sequence is present (read playbook → execute it → follow cog.md, plus a substitution rule for agent delegation), but the operational workflow for the skill's stated task is undefined here and no validation checkpoints exist. This sits at anchor 2 (rough sequence, many gaps, validation absent) rather than 3, since the listed steps are pointer-level directions with none of the actual edit/detect work defined.

Progressive Disclosure

The body is a pure pointer chain: this SKILL.md → `.claude/skills/no-ai-slop/SKILL.md` (the 'authoritative playbook') → `.claude/agents/<name>` workers (the second hop is documented in the body itself), plus `.agents/rules/cog.md`. None of the referenced files exist in the skill's bundle (no references/, scripts/, or assets/ directories), so navigation dead-ends. This matches anchor 1's 'deeply nested references (2+ levels)' pattern; it is not 2 because the problem is not buried references or inlined content but total delegation with 2+ levels of indirection.

Specificity

Two concrete, qualified actions are named — 'Edit drafts into sharper, more human writing while preserving the writer's personal voice' and 'detect AI-slop patterns without rewriting' — which matches anchor 4 (several specific actions, minor gaps). It is not 5 because only two core actions are listed rather than a comprehensive set, and not 3 because both actions are concrete with a clear dual-mode distinction.

Completeness

It explicitly answers both questions: what ('Edit drafts into sharper, more human writing while preserving the writer's personal voice, or detect AI-slop patterns without rewriting') and when ('Use when the user wants a draft clearer, more direct, more opinionated, or less AI-sounding, or asks whether writing reads as AI'). This matches anchor 5's structure of concrete capabilities plus explicit trigger phrases.

Trigger Term Quality

Natural user phrasings comprehensively cover both modes: 'draft', 'clearer', 'more direct', 'more opinionated', 'less AI-sounding', 'reads as AI', 'human writing', 'AI-slop'. These are the phrases a user would actually say ('make this less AI-sounding', 'does this read as AI?'), with synonyms covered; no file extensions apply to this domain.

Distinctiveness Conflict Risk

The AI-slop niche is distinct ('AI-slop patterns', 'AI-sounding', 'reads as AI'), but the editing triggers 'clearer, more direct, more opinionated' could also fire a general editing or proofreading skill. This matches anchor 4 (mostly distinct, minor overlap risk with closely related skills) rather than 5.