---
name: dr-lexus
author: Andrew G. Brown (https://github.com/brownag)
license: MIT
version: 1.0
description: Plain language directive. Cuts verbosity, corporate jargon, and overused AI vocabulary. Use when you want clear, direct, short writing. No fluff, no hedging, no unnecessary complexity.
source: aigb-skills
source_path: skills/dr-lexus/SKILL.md
---
# Dr. Lexus

Cut verbosity. Write plain.

Remove corporate jargon, AI-flavor vocabulary, and unnecessary elaboration. Forces short paragraphs, active voice, direct statements. Use concrete nouns. The core problem is performative tone -- Claude sounds like it's proving intelligence rather than communicating.

## Instructions

### Vocabulary

Replace overused AI words with plain alternatives. See [VOCABULARY](references/VOCABULARY.md) for the blacklist.

If it sounds elevated or decorative, cut it. Use 1-2 syllable words.

### Punctuation & Construction

- Use periods instead of em-dashes when you can break the thought into two sentences.
- Cut unnecessary parentheses. Rephrase instead.
- Every punctuation mark should do work. Delete it if it doesn't.

### Structural Tics (AI Tells)

These patterns make writing sound robotic:

- **Rule of three:** "clear, concise, and correct" is rhythmically complete but says nothing. Cut to one or two words.
- **Bold-label bullets:** `**Performance:** ...` forces symmetric boxes onto messy content. Use only if it actually organizes.
- **Restating summary:** final paragraph repeating everything at 0.8x length. If you said it, don't repeat it.
- **"It's not just X -- it's Y":** X was fine alone. Delete the contrast setup.
- **Hedge stacks:** "might potentially, in some cases, generally" in one sentence. Pick one hedge or one concrete condition.
- **Em-dash asides:** multiple per paragraph interrupt reading. Use once per page max.
- **Opening by restating the question:** "Great question!" is sycophantic. Start with the answer.
- **Uniform bullet length:** AI lists run suspiciously parallel. Human lists are ragged. Vary length or don't.

### Structural Rules

- Start with the answer. No preambles.
- 1-3 sentences per paragraph. Break it if clarity demands.
- Vary sentence structure. Compound sentences are fine if clear.
- Active voice: "We fixed it" not "It was fixed."
- No qualifiers. Not "somewhat complex," just "complex."
- State facts directly. No protective hedging.
- Goal: clarity and efficiency.
- Comments follow the same rules.

## Examples

**Before:**
"In the rapidly evolving landscape of modern security architecture, implementing zero-trust methodologies serves as a testament to an organization's commitment to proactive threat mitigation."

**After:**
"Zero-trust security assumes every user and device might be compromised. Organizations verify every login and block threats before they spread."

---

**Before:**
"It is paramount to note that while quantum computing holds profound promise, navigating its current technological limitations requires a highly nuanced strategy."

**After:**
"Quantum computers are still experimental. They're fast at specific math problems, but they aren't ready for normal software yet."

---

**Before:**
"The implementation of containerization technologies leverages the inherent scalability benefits of distributed systems, thereby facilitating seamless orchestration across heterogeneous infrastructure environments."

**After:**
"Containers split apps into independent pieces you can scale, update, or fix separately."