Conciseness

The ~28-line body is lean with every sentence carrying non-obvious, project-specific knowledge ('The catalog establishes what the installed runtime can actually construct', 'Do not encode substitutions as string rules', 'A partial candidate is never permission to substitute'). No concepts Claude already knows are re-explained and nothing is padded, matching anchor 5: every token earns its place.

Actionability

Guidance is concrete and executable: specific tool invocations with parameters and paths (`cad.catalog` with `action="search"`/`"detail"`/`"acquire"`, `cad.edit` `action="import"`, `cad/research/parts/<wave>/*.json`, `cadPartPlanSchema`) plus a defined exact/partial/rejected taxonomy. It stops short of anchor 5 only because no example payload or report structure is given and 'structured standard filters' is left unspecified — minor gaps, consistent with anchor 4.

Workflow Clarity

Two clearly sequenced numbered workflows (research steps 1-6, main-agent handoff steps 1-4) with real checkpoints: candidate classification (step 5), the unresolved-item rule ('report it as unresolved and state the smallest concrete choice the user must make'), and preservation of unresolved items in the part plan. It is anchor 4 rather than 5 because there is no explicit validate-and-verify step (e.g., confirming the report against its schema) and no error-recovery loop around the acquire/import step.

Progressive Disclosure

The skill is under 50 lines, no bundle files (references/, scripts/, assets/) exist, and none are needed; the body is organized into two well-labeled sections ('## Research workflow', '## Main-agent handoff') with no dangling or buried references. Per the under-50-lines guideline this matches anchor 5: well-organized sections with no need for external references.

Specificity

The description names the domain ('CAD Part Research', 'mechanical components') but the only verbs are the trigger ('Use for') and a generic condition ('must be resolved before modeling') — no concrete actions like search, compare, or report are explicitly stated. It fits anchor 2 (names the domain but actions are minimal or generic), not 3, because no 1-2 concrete capabilities are listed.

Completeness

Both parts are present: an explicit 'when' clause ('when standard, purchasable, or library-backed mechanical components must be resolved before modeling') and a 'what' via the domain label ('CAD Part Research'). It is not 5 because the 'what' names the activity without stating any concrete capabilities; it is not 3 because the 'when' is explicit and specific rather than missing or weakly implied.

Trigger Term Quality

Relevant domain keywords are present ('CAD Part Research', 'mechanical components', 'standard, purchasable, or library-backed', 'modeling') but common variations users would naturally say are missing (e.g., 'part selection', 'catalog', 'off-the-shelf', specific part types like fasteners or bearings). This matches anchor 3: some relevant keywords but missing common variations or synonyms.

Distinctiveness Conflict Risk

The trigger conditions ('standard, purchasable, or library-backed mechanical components must be resolved before modeling') carve out a clear niche that is clearly distinguishable from general CAD modeling or geometry skills, with minimal conflict risk. Anchor 5 fits: clear niche with distinct triggers.