Purpose
Transform technical or data-heavy source text into clear, readable copy calibrated for a specific audience — without dropping, softening, or distorting any factual claim. Solves the recurring problem where subject-matter accuracy and reader comprehension pull in opposite directions, leaving copy that's either correct-but-impenetrable or readable-but-loose.
When to Use
Invoke this skill when the user asks to:
- Rewrite technical documentation or engineering copy for a non-technical audience
- Simplify a data report, research summary, or analytics finding into plain narrative
- Translate jargon-heavy product copy into language customers actually understand
- Make a dense whitepaper, spec sheet, or compliance text accessible without losing precision
- Adapt expert-level content for executives, prospects, or a general readership
- Clean up copy that's accurate but too complicated for its intended reader
Inputs Required
Before starting, collect:
- Source text: The full passage to rewrite (paste the actual copy, not a summary)
- Target audience: Who will read this — e.g., executives, prospects, general public, new employees
- Reader's baseline knowledge: What the audience already knows and doesn't know about the subject
- Purpose of the copy: Inform, persuade, instruct, or reassure
- Non-negotiable claims: Any facts, numbers, or terms that must be preserved exactly
- Length constraint: Target word count or "as concise as possible while complete"
- Tone: Formal / conversational / confident / neutral-institutional
If reader baseline or purpose is missing, assume an informed-but-non-expert reader whose goal is to understand and act, and note the assumption.
Instructions
Step 1: Extract and Lock the Factual Claims
Before rewriting a single sentence, inventory what must survive.
- List every distinct factual claim in the source: statistics, causal statements, technical mechanisms, conditions, caveats, and defined terms
- Flag any claim that is qualified ("up to," "in most cases," "under X conditions") — qualifiers are load-bearing and cannot be silently dropped
- Mark terms that are genuinely technical (must be retained or explicitly defined) vs. jargon that has a plain-language equivalent
- This inventory becomes the accuracy contract: the rewrite must honor every locked claim
Step 2: Diagnose the Complexity Sources
Identify why the source is hard to read — the fix depends on the cause:
- Lexical complexity: Jargon, acronyms, nominalizations ("utilization of" → "using")
- Syntactic complexity: Long sentences, nested clauses, passive voice hiding the actor
- Structural complexity: No hierarchy — everything presented as equally important
- Cognitive complexity: Abstract concepts with no concrete anchor; data with no interpretation
- Density: Too many claims per sentence, forcing the reader to unpack multiple ideas at once
Step 3: Apply the Simplification Moves
Rewrite using targeted techniques, matched to the diagnosis:
- Unpack density: One idea per sentence. Split any sentence carrying two or more distinct claims.
- Concrete over abstract: Replace abstractions with tangible examples, comparisons, or analogies the audience already understands — without introducing a claim not in the source
- Active voice, named actors: Say who does what. Convert "errors are reduced" → "the system reduces errors"
- Contextualize data: Never leave a number bare. Give it meaning ("40% faster — that's ten minutes saved per report")
- Define, don't drop: Retain a technical term only if the audience needs it; then define it inline on first use
- Build a hierarchy: Lead with the takeaway, then support it — invert the source if it buries the point
- Cut, don't dumb down: Remove filler, hedging, and repetition; simplifying is subtraction of noise, not of substance
Step 4: Preserve Credibility While Simplifying
Simple copy can read as vague or overstated — guard against both:
- Keep every qualifier from the claim inventory ("up to," "typically," "in tested conditions")
- Do not upgrade a hedged claim into an absolute one ("can help reduce" ≠ "reduces")
- Retain the strongest verifiable specifics — precise numbers and named mechanisms build trust more than adjectives
- Avoid replacing a defensible technical statement with a marketing claim that the source doesn't support
Step 5: Verify Against the Accuracy Contract
Cross-check the rewrite line by line against the Step 1 inventory:
- Confirm each locked claim appears, with its qualifiers intact
- Confirm no new claim was invented during simplification
- Confirm no number, condition, or causal relationship changed meaning
- Flag any claim you could not preserve without distortion, and note it rather than dropping it silently
Step 6: Score Readability
Assess the rewrite on measurable dimensions:
- Estimated reading grade level (target: appropriate to the audience — typically grade 7–9 for general, 9–11 for professional)
- Average sentence length (target under 20 words)
- Jargon count remaining (and whether each remaining term is defined)
- Passive-voice instances
Output Format
- Structure: Four labeled sections in order
- Length: Rewrite matches the requested length; full deliverable typically 600–1,000 words including analysis
- Format: Rewritten copy as clean prose; before/after as a two-column-style comparison; claims log and scorecard as bulleted or tabular lists
- Sections: 1. Rewritten Copy — the finished, deliverable-ready text 2. Before / After Comparison — 3–5 representative sentence pairs showing the source and the rewrite side by side 3. Flagged-Claims Log — every preserved claim confirmed, plus any claim that could not be simplified without risk (with explanation) 4. Readability Scorecard — grade level, avg sentence length, remaining jargon (with definition status), passive-voice count
- Tone: The rewrite matches the requested audience tone; the analysis sections are direct and editorial
Quality Checklist
Before delivering output, verify:
- [ ] Every claim from the Step 1 inventory appears in the rewrite, with its qualifiers intact
- [ ] No new factual claim was introduced that the source did not support
- [ ] No number, condition, or causal statement changed meaning between versions
- [ ] Average sentence length in the rewrite is under 20 words
- [ ] Every technical term retained is defined on first use; every droppable jargon term was replaced
- [ ] The Flagged-Claims Log explicitly names any claim that could not be preserved cleanly — nothing was dropped silently
Notes
- The single most common failure mode is upgrading a hedged claim into an absolute one to make the copy punchier — this trades credibility for polish and often creates legal or factual exposure. Preserve the hedge.
- Data almost always needs interpretation, not just translation: a number the reader can't contextualize is noise, so pair every retained statistic with what it means for them.
- When the audience genuinely needs a technical term (a doctor reading clinical copy, an engineer reading a spec), simplify the surrounding language but keep the term — simplification is about the reader's effort, not a fixed vocabulary list.
- If the source contains a claim that is unclear or possibly inaccurate in the original, flag it rather than smoothing it over — a rewrite is not the place to launder a shaky assertion.
How to use this skill
This file is a Claude Code skill in Anthropic's standard format — markdown with YAML frontmatter at the top.
Installing in Claude Code:
~/.claude/skills/<folder-name>/SKILL.md— auto-loads in every session anywhere on your machine (recommended for skills you'll use across multiple projects)<your-project>/.claude/skills/<folder-name>/SKILL.md— auto-loads only when working in that project (recommended for project-specific skills)
Restart Claude Code or open a new session to register.
Using in Claude.ai, ChatGPT, Gemini, or other AI tools:
- Drag this file into a new chat, or copy and paste its contents
- Send your task or request as your next message — the AI will apply the skill where it fits
For Claude Code skill details: https://docs.anthropic.com/en/docs/claude-code/skills