Conciseness

The body is lean and operational throughout — e.g. "Pass the resolved root ID to every content call. Omit specNodeId only when the file is known to contain exactly one Figma spec; omission throws otherwise" packs exact behavioral rules into single sentences. There is no padding, no explanation of concepts Claude already knows (GFM is mentioned only as a factual format note), and every token carries non-obvious API-specific information. Matches the score-5 anchor.

Actionability

Every operation has copy-paste-ready executable code with real arguments and error checks: figma.spec.createOnCanvasAsync, getContent/setContent/editContent, comments.create/getThread/reply, including post-write verification ("if (!after.markdown.includes('A paragraph.')) throw new Error(...)"). Matches the fully-executable score-5 anchor.

Workflow Clarity

The sequence is explicit with validation checkpoints throughout: the mandatory whoami + specs_enabled gate up front ("If specs_enabled is false or absent for that plan, explain that Specs is unavailable and stop the workflow"), the page-by-page discovery loop, getContent-before/after verification for every edit ("DON'T rely on screenshots"), and explicit failure modes for destructive ops ("editContent requires exactly one match. Zero or multiple oldText matches fail without changes"). This matches the score-5 anchor with feedback loops for destructive document operations; not 4 because no validation step is missing or implicit.

Progressive Disclosure

No bundle files exist (references/, scripts/, assets/ are all absent), and the single file is well organized into clearly headed sections (Discover, Create, Content methods, Critical rules, Style, Images, Comments) with all content being core operational guidance. This lands at 4 rather than 5 because at ~113 lines the Comments and Images sections are self-contained detail blocks that could alternatively live in one-level-deep reference files, and the file is above the under-50-lines threshold the simple-skill exception covers. It is well above 3 since nothing is buried and no content that belongs elsewhere is inlined to the point of harming navigation.

Specificity

"This skill is required to create, read, edit, review, or comment on Figma specs" lists five concrete actions, which initially matches the score-4 anchor (several specific actions, minor coverage gaps). However, the description shifts to second person ("You must read it when working on..."), and the judging guidelines mandate a 1-point specificity reduction for second-person voice, lowering it to 3. It is not a 2 because the capability list itself is concrete and multi-action, not minimal or generic.

Completeness

The "what" is explicit ("create, read, edit, review, or comment on Figma specs") and the "when" is explicit and concrete ("You must read it when working on a PRD, brief, RFC or any kind of document in either a standalone Figma spec or a Figma spec embedded in a Figma Design or Figjam file"). Both are answered clearly with concrete trigger phrases, matching the score-5 anchor.

Trigger Term Quality

Natural trigger phrases are present: "PRD, brief, RFC", "Figma spec", "standalone Figma spec", "Figma Design or Figjam file" — terms a user would plausibly say. It is not 5 because common synonyms such as "specification", "product spec", or "Figma file" variations are missing; it is above 3 because several distinct, naturally-spoken keywords are covered.

Distinctiveness Conflict Risk

"Figma spec", "Figjam", and "Figma Design" carve out a clear niche with distinct, product-specific triggers unlikely to collide with generic document or PDF skills. The broad phrase "any kind of document" is tightly scoped to Figma spec contexts, so conflict risk is minimal.