Conciseness

Nearly every token is non-inferable, API-specific knowledge (auth header, endpoint, parameter shapes, status semantics, topUpTo floors, unresolvedPrograms) with no padding of concepts Claude already knows. However, several specific dates ('scheduled for 2026-11-10', 'not collected since 2026-02-24') sit in the main included/not-included flow rather than a labeled time-sensitive section, which the rubric penalizes — efficient but with trimmable instances, so level 4 rather than 5.

Actionability

Fully executable, copy-paste-ready: a complete curl command with --data-urlencode and a jq projection, a worked example with real numbers (16.5% MFN plus 10% under 9903.05.44 = 26.5%), a full parameter table, and enumerated error codes. The common case is completely covered.

Workflow Clarity

A single-call skill whose action is unambiguous, with explicit validation checkpoints: read each duty's status before summing, check estimateComplete and additionalDutiesLoaded before quoting a total, note upstreamUnavailable, and handle 401/403/400/429 with retry-with-backoff guidance. Under the simple-skill exception this earns the top score.

Progressive Disclosure

Self-contained skill with no bundle files and clear, well-labeled sections (Authentication, Endpoint, Parameters, Response shape, Worked example, Errors). The dense response-semantics block (status/basis/estimatedRate paragraphs) is reference-like material inlined in SKILL.md, a minor organization gap that keeps it below the well-split top anchor but above the 'should be separate is inline' level 3.

Specificity

The single action is comprehensively and concretely specified: 'Look up the current US import duty' returning 'the HTS column rate (MFN, trade-agreement or column 2) plus the Section 301 and 232 duties in chapter 99' from 'the live USITC Harmonized Tariff Schedule'. No coverage gaps in what is returned, so it matches the comprehensive anchor rather than the 'minor gaps' level 4.

Completeness

Explicitly answers both questions: what ('look up the current US import duty... from the live USITC Harmonized Tariff Schedule') and when ('Use when the user asks what the US charges today on a product from a specific country, including China tariffs'). Both are concrete and explicit with trigger phrases, matching the top anchor rather than the 'when could be more explicit' level 4.

Trigger Term Quality

Good natural phrases users would actually say: 'what the US charges today on a product from a specific country', 'China tariffs', 'US import duty', 'additional duties'. A few natural synonyms are missing ('customs', 'customs duty', an unqualified 'tariff rate'), placing it at the 'good coverage, a few terms missing' anchor rather than comprehensive.

Distinctiveness Conflict Risk

Clear niche with distinct triggers — US import duties on HTS codes from a specific partner country, chapter 99 actions, USITC HTS as source. This would not plausibly fire for any other skill; overlap risk is minimal.