Conciseness

Nearly every token is tool-specific knowledge Claude cannot infer — real commands, flag semantics, and gotchas like "Roughly 60% of references are not calls" and may-semantics — with no explanations of concepts Claude already knows. Not 5: there are minor trims available, e.g. the duplicated example `webspec-index query 'DOM#concept-tree'` and the pseudo-syntax placeholder line in the state section. Not 3: no padded or unnecessary explanatory sections.

Actionability

Every section is copy-paste-ready bash with realistic concrete arguments ('HTML#navigate', "RFC9000#section-22", 'HTML#*sandbox*', 'Window.navigation'), documented flags, and described outputs. Specific examples cover the common cases for each subcommand, matching the fully-executable score-5 anchor.

Workflow Clarity

Scenario workflows sequence commands with comments, and validation checkpoints with error-recovery loops are explicit: "Read `Search was truncated` before trusting the result — raise `-l` or narrow the endpoints", "confirm an absence with `refs --direction incoming --kind step` before concluding a call does not exist", and `exists` exit-code validation before citing anchors. The tool involves no destructive or batch operations, so no cap applies.

Progressive Disclosure

No bundle files exist (references/, scripts/, assets/ are absent) and all ~440 lines — including the full per-command API reference — are inlined in SKILL.md. Section headers give it some structure, but the command-reference bulk clearly belongs in a one-level-deep references/ file, matching the score-3 anchor (content that should be separate is inline) rather than 4 (most content appropriately placed).

Specificity

The description enumerates eight concrete, distinct actions ("read a spec section or algorithm", "find which algorithms invoke a concept and with what arguments", "trace a call chain", "list who writes a spec field", "look up IDL", "validate an anchor", "read a spec as changed by an open WHATWG PR") — comprehensive coverage of the tool's capabilities with no vague filler. It clearly matches the score-5 anchor rather than 4, which would require minor gaps in the action inventory.

Completeness

It explicitly answers both questions: the "what" is the opening sentence ("Query WHATWG, W3C, IETF and TC39 web specifications from the command line"), and the "when" is an explicit "Use to…" clause with concrete trigger phrases for each capability. Not 4: the trigger guidance is already explicit and specific, not merely present.

Trigger Term Quality

It uses natural phrases a user would actually say when needing this skill ("spec section", "algorithm", "call chain", "IDL", "anchor", "PR", plus the org names WHATWG/W3C/IETF/TC39), giving good keyword coverage. It falls short of 5 because common synonyms like "standard", "web standard", or "spec URL" are missing.

Distinctiveness Conflict Risk

Naming the four standards bodies and the command-line query niche gives it a clear domain with distinct triggers; no plausible overlap with unrelated skills. Minimal conflict risk, matching the score-5 anchor.