The body is lean: compact tables for issue detection, no re-explanation of concepts Claude already knows, and every line earns its place; not 2 because there is no padding or unnecessary explanation to tighten.

Provides concrete copy-paste fix formulas ('=TRIM(A2)', '=VALUE(SUBSTITUTE(B2,"$",""))', '=UPPER(C2)', '=DATEVALUE(D2)') and an Office JS pattern, but the profiling/detection step ('Profile each column: detect its dominant type') and the empty propose-fixes table lack executable code; not 3 because it is not fully copy-paste ready across all steps, not 1 because substantial concrete guidance is present.

A clear 4-step sequence (Scope → Detect → Propose → Apply) with explicit validation checkpoints ('Show a summary table before changing anything', 'confirm with the user first', 'show the user a sample... and get confirmation before moving to the next category'); the destructive/batch feedback loops satisfy the top anchor.

Under 50 lines with no external references needed and well-organized sections (Environment, Workflow, tables), meeting the simple-skills allowance for a top score without bundle files.

Lists multiple specific concrete actions — 'trim whitespace, fix inconsistent casing, convert numbers-stored-as-text, standardize dates, remove duplicates, and flag mixed-type columns' — matching the top anchor; not 2 because it goes well beyond naming a domain plus a few actions.

Clearly answers both what (the cleaning actions) and when ('Use when data is messy, inconsistent, or needs prep before analysis') with explicit triggers, satisfying the top anchor.

Explicit natural trigger phrases ('clean this data', 'clean up this sheet', 'normalize this data', 'fix formatting', 'dedupe', 'standardize this column', 'this data is messy') give good coverage of terms users would actually say.

Scoped to spreadsheet data cleaning with distinct triggers, giving it a clear niche unlikely to fire for unrelated skills; not 2 because the triggers are specific rather than broadly overlapping.