content-polishlisted

# Content Polish: End-to-End Workflow This skill orchestrates the **humanizer** and **proofreader** skills in the correct order, with proper handoff between them, and adds a critical middle step that catches the most common failure mode: **over-fragmentation** (content reading like a list of short sentences instead of connected prose). ## Why This Skill Exists Running humanizer then proofreader sequentially as separate skills produces sub-par results because: 1. The humanizer can over-correct: it strips em-dashes, kills -ing phrases, and mechanically "varies rhythm" which fragments connected ideas into staccato sentences. 2. The proofreader, applied to a CV, doesn't know that parallel structure across bullets is the format, not an error. 3. Neither skill knows the other ran. Format conventions get lost in handoff. 4. There's no anti-fragmentation pass between them so over-chopped content goes into the proofreader, which often fails to catch the issue (or worse, suggests further "clarity" fixes that fragment more). This skill fixes all four problems. --- ## The Workflow ### Step 0 - Identify Document Type (DO THIS FIRST, ALOUD) State the document type explicitly to the user before starting. This is the most important calibration decision and everything downstream depends on it. Possible types: - **CV / Resume** (bullets, parallel structure, action verbs) - **Cover letter** (first-person warm-professional) - **LinkedIn post** (mixed rhythm, opinion ok) - **Blog post /