How Do You Make Claude Opus 5 Clearer and Shorter? Anthropic’s Fix

TL;DR

Make Claude Opus 5 clearer by creating a custom ELI5 output style in Claude Code and adding ASD-STE100 Simplified Technical English rules to reduce jargon. Because output styles enter the core system prompt and receive reminders during sessions, they can be more consistent than instructions in Claude.md. For occasional walls of text, use a separate skill to shorten the last reply. Read on for the setup and confirmation steps.

Transcript

There's a widespread problem plaguing Claude users recently, specifically on Opus 5, because its replies [music] tend to be confusingly long, packed with all too possible nonsense, and it's becoming more and more jargon dense. The good news is that Anthropic already shared a fix for this. It's super easy, and the principle behind it makes your setu... Read More

Key Insights

• Claude Opus 5 replies can be harder to use because they often default to comprehensive explanations, technical terminology, unexplained acronyms, and long blocks of text, even when the user asks a relatively simple question.
• Long Claude responses impose two costs: users must spend more time reading and interpreting the answer, while the additional output also consumes more tokens. A useful configuration should therefore improve clarity without permanently removing detail needed for demanding work.
• Claude Code's output style controls how the system communicates its answers. The built-in configuration includes default, proactive, explanatory, and learning options, while users can add a custom style such as ELI5 for simpler language.
• ELI5 means explain like I am five and directs Claude toward language that is easier to understand. The demonstrated custom style leads with the central answer, tells a clearer story, and closes with an understandable conclusion.
• ASD-STE100 Simplified Technical English is presented as a controlled-language standard with a more restricted, accessible vocabulary. Adding it to the ELI5 style can reduce jargon, vague technical wording, and unnecessary complexity in conversations with agents.
• A custom output style can be installed by asking Claude Code to save the supplied instructions in its output styles folder and update the appropriate settings files. Starting a new session and asking which style is active provides a confirmation check.
• Output-style instructions can be more consistent than Claude.md instructions because the style becomes part of Claude Code's core system prompt. Claude Code also injects reminders during the session to keep responses aligned with the selected style.
• A separate skill is better suited to occasional wall-of-text problems because some tasks genuinely require detailed answers. The demonstrated slash command restates Claude's previous message in plain human language with zero jargon only when the user invokes it.

Questions & Answers

Q: How can I make Claude Opus 5 clearer and use less jargon?

Create a custom ELI5 output style in Claude Code and add an instruction to follow ASD-STE100 Simplified Technical English. This directs Claude toward simpler vocabulary, fewer unexplained technical terms, and answers that lead with the central point.

Q: What is an output style in Claude Code?

An output style controls how Claude Code presents information and communicates with you. Claude Code includes default, proactive, explanatory, and learning options, and it also supports custom styles such as ELI5.

Q: How do I install a custom ELI5 output style in Claude Code?

Give Claude Code the ELI5 instructions and ask it to save them in the output styles folder and update the appropriate settings files. Then begin a new session and ask which output style is active to confirm that the installation worked.

Q: Why use an output style instead of putting the rules in Claude.md?

Output-style rules become part of Claude Code’s core system prompt. Claude Code also adds reminders during the session to follow the selected style, which can make these rules more consistent than instructions placed only in Claude.md.

Q: What does ELI5 mean, and how does it change Claude’s answers?

ELI5 means “explain like I am five.” The custom style encourages Claude to begin with the main answer, explain the subject in easier language, and finish with a conclusion that does not require specialist knowledge.

Q: What is ASD-STE100 Simplified Technical English used for?

ASD-STE100 Simplified Technical English is presented as a controlled-language standard that uses a more restricted and accessible vocabulary. Adding its rules to the ELI5 style can reduce jargon, vague technical language, and unnecessary complexity.

Q: How can I shorten a wall-of-text response from Claude Opus 5?

Use a dedicated skill that tells Claude to restate its previous message in plain human language with zero jargon. Invoking the skill only when needed produces a more direct version without forcing every response to be short.

Q: Why should concision be handled by a separate skill?

Some knowledge-work and production-code tasks genuinely require detailed answers. Keeping concision in an optional skill lets you shorten a specific response while leaving the permanent ELI5 output style focused on clarity and reduced jargon.

Summary & Key Takeaways

• Claude Opus 5 can produce answers that are unusually long, jargon-heavy, and filled with plausible-sounding material that requires extra effort to evaluate. This creates two practical problems: users spend more time finding the main point, and longer responses consume more output tokens, increasing the cost of working with Claude Code.

• The recommended jargon fix is a custom ELI5 output style, optionally strengthened with ASD-STE100 Simplified Technical English. Users can ask Claude Code to create the style file, save it in the output styles folder, update the relevant settings, and then begin a new session to confirm that the style is active.

• Output-style rules are presented as more dependable than equivalent instructions in Claude.md because they enter Claude Code's core system prompt and receive additional reminders during a session. Concision should remain separate and optional, since detailed output is sometimes necessary for knowledge work or production code. A skill can shorten replies when needed.