Conciseness

The body is dense, non-inferable specification — exact token values, the scaling formula `factor = min(innerWidth/1920, innerHeight/1080)`, exact class definitions, a CSS gotcha — with essentially no explanation of concepts Claude already knows. It falls short of a 5 only because the "copy verbatim / do not rewrite / keep the seed's script verbatim" enforcement is repeated 3-4 times across sections and a few flourishes ("Sticker energy", "a retro poster shop wired to a power grid") could be trimmed.

Actionability

Fully executable guidance: a copy-paste-ready `:root` token block, the exact Google Fonts URL, exact halftone `radial-gradient` values, concrete class names (`.panel-blue`, `.badge`, `.ghost-num`), precise stage-scaling math, a layout-masters table mapping each master to a seed slide, and a hard output contract. This is well beyond anchor 4's "concrete code with minor gaps" — specific, complete, and copy-paste ready.

Workflow Clarity

The opening paragraph gives an unambiguous core flow ("Start from `example.html`... copy its stage CSS, slide shell chrome, signature-device CSS, and the entire `SlidePresentation` controller script verbatim, then replace only the slide content"), followed by logically ordered spec sections and explicit constraints (output contract, forbidden lists). Not a 5 because there is no explicit validation/verification step (e.g., check overflow/overlap at 1920×1080 before finishing) — the critic pass appears only in the frontmatter example prompt, not in the body.

Progressive Disclosure

The body is well-sectioned, but it is a ~109-line monolithic spec whose central instruction depends on `example.html`, `STYLE_PRESETS.md`, and a `LICENSE` file — none of which are present in the provided bundle (no references/, scripts/, or assets/ directories, and no example.html alongside SKILL.md). Scoring against the actual bundle structure, the clearly-signaled references dangle unverified, and the seed CSS/JS plus bulk spec could be split into reference files. More structured than anchor 2, but the missing primary reference target keeps it below anchor 4.

Specificity

The description is a topic statement ("OpenDesign's seed pitch: the open, local alternative to closed AI design — why now, the wedge, and the ask. Built as a decision-grade fundraising pitch deck for pre-seed & seed VCs") — it names the fundraising-deck domain but contains no action verbs describing what the skill does. It is concrete about the subject matter, so not a 1, but has zero stated skill actions, so it cannot reach 3.

Completeness

The "what" is vague and misdirected — it describes one deck's content rather than the skill's capability — and there is no "Use when..." clause or equivalent trigger guidance, so the cap applies regardless. This matches anchor 2 ("vague 'what' and no 'when'"); it cannot reach 3 because the what does not state the skill's function at all.

Trigger Term Quality

"seed pitch", "fundraising pitch deck", and "pre-seed & seed VCs" are natural phrases, but common variations users would say ("investor deck", "slides", "presentation", standalone "fundraising") are missing, and much of the wording ("closed AI design", "the wedge", "the ask") is example-topic vocabulary rather than trigger terms. Coverage is broader than anchor 2 but not the near-complete coverage of anchor 4.

Distinctiveness Conflict Risk

The seed-VC fundraising niche is somewhat specific, but "pitch deck" overlaps with general presentation/deck skills and the description offers no functional hooks (themed HTML deck generation) to distinguish it. More specific than anchor 2's broad document handling, but not as clearly distinct as anchor 4.