logophile_skill

Overview

This skill edits copy for clarity and semantic density, tightening wording while preserving intent and tone. It compresses verbose text, sharpens names and labels, and produces faster-to-scan output. Default mode returns the revised text only; optional modes add inline edits or reduction deltas.

How this skill works

On request, it distills the source into a one-sentence intent and preserves required tokens (numbers, code, quotes). It removes filler, verbifies nominalizations, and applies a precision ladder to replace vague words with exact actions or properties. The output is reshaped to lead with actions, keep sentences atomic, and parallelize lists; verification ensures obligations, risks, and facts remain unchanged.

When to use it

• Tighten long emails, docs, prompts, or specs that are slow to scan
• Compress repetitive or filler-heavy paragraphs without losing meaning
• Refine names, titles, labels, or headings for clarity and impact
• Sharpen prompts for LLMs to increase signal per token
• Produce compact versions for UIs with space constraints

Best practices

• Provide must_keep tokens (numbers, code, exact phrasing) when critical
• Specify desired tone or length target if different from the source
• Use the default fast mode for quick rewrites, annotated for review
• Avoid asking for creative rewording that would change technical meaning
• Request format constraints (bullets, table, single sentence) up front

Example use cases

• Shrink a three-paragraph feature spec into a one-paragraph summary while keeping acceptance criteria
• Rewrite onboarding UI labels and headings to fit mobile space without losing clarity
• Condense a long support email into a concise response that preserves required steps and deadlines
• Compress a verbose prompt into a token-efficient version for cost-sensitive model calls
• Generate multiple short title options from a long article headline