Natural Language Editor

Rewrite user-provided text to sound natural, clear, and smooth without changing meaning or factual content. Use when polishing drafts, removing robotic phras...

Security Scan

Instruction-only text-rewriting skill that is internally consistent with its stated purpose and requests no extra permissions, installs, or credentials.

Purpose & Capability

Name and description (natural-language editing) match the SKILL.md commands and rules. The skill asks for no binaries, env vars, or installs — appropriate for a pure text-editing instruction set.

Instruction Scope

SKILL.md only describes how to rewrite provided text (modes, constraints, and output formats). It does not instruct the agent to read unrelated files, access environment variables, call external endpoints, or exfiltrate data.

Install Mechanism

There is no install spec and no code files to write or execute. As an instruction-only skill, it has minimal surface area and does not perform on-disk installation.

Credentials

The skill requires no environment variables, credentials, or config paths. Requested capabilities are proportional to a text-editing skill.

Persistence & Privilege

always is false and the skill does not request persistent or elevated privileges or modification of other skills. Autonomous invocation is the platform default and not a special request here.

Assessment

This skill appears coherent and low-risk: it's an instruction-only text rewriter with no external installs or secret access. Before using it, avoid submitting sensitive personal, proprietary, or regulated data (the agent/model processes any text you send). As a best practice, test the skill on non-sensitive samples to verify it truly preserves facts and does not introduce undesired claims, and confirm you are comfortable with the platform/model handling the text.

Rewrite user text so it reads naturally while preserving intent and facts.

Commands

• edit [text] — default natural rewrite
• edit --concise [text] — shorter, same meaning
• edit --clarity [text] — easier to understand
• edit --professional [text] — neutral, polished tone
• edit --light [text] — minimal edits only
• edit --deep [text] — stronger restructuring, same meaning
• edit --anti-ai [text] — reduce templated/AI-like cadence while preserving meaning
• edit --linkedin [text] — optimize for LinkedIn profile/about style (credible, concrete, concise)
• edit --keep-length [text] — preserve source length within ±10% while improving naturalness
• edit --json [text] — structured analysis + rewrite

If the user provides only text, use default natural mode.

Core Rules

1. Preserve meaning.
2. Do not add new facts.
3. Do not remove essential claims.
4. Do not inject new opinions, emotions, or experiences.
5. Keep perspective unless the user asks to change it.
6. Prefer direct, concrete wording over vague phrasing.
7. Keep technical terminology accurate when precision matters.

Mode Behavior

Natural (default)

Improve flow and readability with minimal semantic change.

Concise

Remove redundancy and filler; keep all essential points.

Clarity

Simplify complex phrasing, split overloaded sentences, reduce abstraction.

Professional

Use clear, neutral, business-appropriate wording.

Light

Keep structure and tone close to source; apply minimal edits.

Deep

Allow stronger sentence-level restructuring while preserving meaning.

Anti-AI

Reduce robotic markers (template transitions, repetitive cadence, hedge stacking, generic intensifiers). Vary sentence rhythm and prefer concrete wording while preserving facts and intent.

LinkedIn

Optimize for LinkedIn profile/about usage: strong first-person positioning, concrete outcomes, credible tone, and scannable structure. Avoid hype, clichés, and overclaiming.

Keep-Length (±10%)

Rewrite naturally while keeping total character length close to the source (target within ±10%). Prioritize meaning preservation first, then length control.

Rewrite Workflow

1. Identify friction points (repetition, filler, generic transitions, vagueness, rhythm issues, tone mismatch, AI-like templating).
2. Choose the least invasive mode that solves the problem.
3. Rewrite while preserving facts, claims, terminology, and perspective.
4. For --anti-ai, explicitly remove robotic patterns and smooth cadence without adding subjectivity.
5. For --linkedin, prioritize concrete achievements, clean role framing, and concise professional tone without adding new claims.
6. For --keep-length, check source vs rewrite length and adjust to stay within ±10%.
7. Validate output: no factual additions, no distortion, better readability.

Output Format

• Default: return rewritten text only.
• JSON mode (--json):

{
  "issues_found": ["repetitive phrasing", "inflated language"],
  "rewrite_strategy": "remove filler and simplify wording",
  "rewritten_text": "...",
  "introduced_subjectivity": false
}

Limits

Do not:

• add new facts,
• change argument intent,
• invent experiences/emotions,
• shift perspective unless requested,
• strengthen legal/medical/technical claims beyond source.

Author

Vassiliy Lakhonin