
AI prompt optimizer & prompt engineering skill. AI 提示词优化、需求澄清。Portable guidance for GPT, DeepSeek
AI prompt optimizer & prompt engineering skill. AI 提示词优化、需求澄清。Portable guidance for GPT, DeepSeek, Claude, Gemini, Qwen, Kimi, GLM, Doubao, Hunyuan, ERNIE, MiniMax, Grok, Mistral & Llama; integrations depend on the host.
·
English · 简体中文 · 繁體中文 · 日本語 · 한국어 · Türkçe
Русский · Tiếng Việt · ไทย · Deutsch · Español · Français
Українська · Polski · Português (Brasil) · Pāli · Bahasa Indonesia · Italiano
Ελληνικά · العربية · हिन्दी · বাংলা · فارسی · עברית
Nederlands · Svenska · Português (Portugal) · Bahasa Melayu · Filipino · Kiswahili
Turn vague requests into clear, actionable AI instructions—while preserving what you meant.
AI Prompt Optimizer is a prompt optimization and requirements clarification Agent Skill. Use it to improve prompts, audit conflicting instructions, or clarify a task before an AI agent executes it. It covers coding, design, writing, analysis and planning.
AI 提示词优化助手:优化 Prompt、澄清需求、明确任务边界与验收标准,保留原意。
v2.4.0 · MIT · Instruction-based skill + optional installer · No API key required by the skill
Get started · User guide · 中文使用指南 · Evaluation record · Publishing guide
Your request
Make this product page feel more premium. Keep all text, prices and button behavior unchanged. It should look good on mobile too.
A clarified instruction
Improve the existing product page's visual presentation while preserving all text, prices and button behavior, including labels and destinations. Provisionally interpret “premium” as clear hierarchy, consistent typography, balanced spacing and restrained styling aligned with the existing brand. Check mobile layouts for overflow, overlap, readability and usable controls. Compare the result against the original content and behavior. Do not add new modules or redesign the brand.
This is an illustrative translation, not a fixed template or performance claim.
| Your situation | What the skill adds |
|---|---|
| “Make it professional” | Context-specific structure, tone or visual criteria. |
| “Be comprehensive” | The relevant workflow, evidence and consequential exceptions. |
| “Make it faster” | A measurable baseline and comparable verification, without invented targets. |
| “Don't ask me lots of questions” | Reversible assumptions for small details; explicit handling of material unknowns. |
| “Just improve my prompt” | A copyable instruction without executing the embedded task. |
| “Clarify it and do it” | Minimal clarification followed by authorized execution. |
The skill preserves numbers, names, exclusions and scope. It separates known information from assumptions, detects conflicting requirements and keeps already-clear tasks short. It includes 34 semantic patterns, 10 domain guides and 8 teaching examples, loaded only when useful.
30 language/locale entries with clickable README and getting-started guides. Added language rules distinguish the explanation, optimized prompt and downstream output; preserve scripts, variables and exact constraints. Pāli has an experimental bilingual starter. Added translations are AI-authored and not professionally native-reviewed; documentation coverage is not a model-quality guarantee.
Language directory and coverage · Language validation
Repair prompts from failed outputs; preserve reusable variables and output contracts; compare evidence without inventing scores; consolidate multi-turn context; refine image/video briefs. Simple requests remain short, and provider calls are not added by default. Use natural language—no new command syntax.
Advanced usage · Research and decisions · New validation record
Deterministic personal/project installation, optional on-demand guidance, status checks, safe updates and backed-up removal. Ask local Codex to install the supplied package. The installer requires Python 3.9+; everyday skill use does not. Native plugin packaging is an unverified preview.
Recommended installation and maintenance · Validation and platform limits
Give Codex this package and send:
Read docs/INSTALL.md, install AI Prompt Optimizer for my personal Codex and enable on-demand refinement. Use the included installer, preserve existing configuration and local edits, and report actual verification.
Clone this repository or use GitHub’s Code → Download ZIP, then supply the extracted folder to Codex. See publication status. For requirements, old-version migration and install-only mode, see the installation guide.
Download ZIP, extract and name the skill directory ai-prompt-optimizer, then place it in ~/.agents/skills/ or the project's .agents/skills/. It must directly contain SKILL.md, agents/ and references/. Manual copying does not configure a default preference. See Git/Windows instructions and existing-installation migration.
In Codex CLI or IDE, invoke the skill by name:
$ai-prompt-optimizer Improve this prompt only: Make my report more professional without changing the facts.If your client provides a skill picker, select AI Prompt Optimizer. If it does not appear, start a new session or restart the client. Automatic selection depends on the host and task; explicit invocation is the most predictable way to use it.
| Mode | Example |
|---|---|
| Refine only | $ai-prompt-optimizer Rewrite this as a clear task instruction. Do not execute: … |
| Clarify and execute | $ai-prompt-optimizer Clarify and then carry out this task: … |
| Audit | $ai-prompt-optimizer Check this prompt for conflicts and missing essentials: … |
No special command parser is required. If execution is not requested, the skill defaults to refinement.
Documented model families: GPT, DeepSeek, Claude, Gemini, Qwen, Doubao, Hunyuan, ERNIE, GLM, Kimi, MiniMax, Grok, Mistral and Llama. The shared instruction workflow is designed for their suitable text/chat variants; individual model integrations have not been tested. See the 14-family usage matrix for official references, usage paths and validation status.
| Environment | What is supported by this project |
|---|---|
| Codex skill-capable environments | Standard SKILL.md structure with optional UI metadata. Local discovery paths follow the official documentation linked in the user guide. |
| Other Agent Skills hosts | Portable Markdown instructions. Check that host's discovery paths, metadata support and available tools. Not separately integration-tested. |
| Listed model families through a suitable client or API | Reusable natural-language guidance. The client must provide context and any tools. Separate model API comparisons have not been run. |
| Plain chat without Skills | Paste the SKILL.md body as guidance, provide the original request separately, and attach only relevant reference material. See the fallback workflow. |
This is a skill, not a hosted web app, model, browser extension or automatic model-settings switch. It cannot create missing tools, bypass permissions, guarantee correctness or eliminate every clarification round.
| English | 简体中文 |
|---|---|
| Model compatibility and usage | 大模型适用与使用方法 |
| Detailed user guide | 详细使用指南 |
| GitHub publishing and discovery | GitHub 发布与搜索配置 |
The language links switch between maintained Markdown documents. They do not change GitHub's interface language or automatically translate content. Core skill instructions are maintained in one file to avoid behavior drift; output follows the user's language.
Eight independent agent tasks were observed in v1.0.0, including a real repair of a synthetic Python function with four existing tests. The core workflow is unchanged through v2.1.0. v2.0.0 renamed the skill; v2.1.0 adds model-family portability documentation without new API integrations. See validation details and limitations. The original suite contains sixteen cases, eight unrun. New v2.2.0 behavior cases and their actual observations are recorded separately in the new evaluation record. These are not cross-model benchmarks or proof of reduced cost or follow-up count.
This project was previously named clarify-intent / 意图精炼器. v2.0.0 changes the invocation to $ai-prompt-optimizer. No automatic alias is installed for the old command. Update saved prompts and local folder names; keep only one active copy. See migration instructions.
Contribute a concrete input, expected behavior and a counterexample. Do not include private data or credentials. See the contribution and evaluation workflow.
MIT License. The original copyright notice is retained. This project is independent and is not an official product of any listed model provider.
This repo does not document an install step we could read. Check the README on GitHub.
Paste this prompt into your agent. It reads this page and does the setup for you.
Read https://aiagentslisting.com/mcp/ai-prompt-optimizer to learn what the "AI Prompt Optimizer" MCP server does and how to install it. Add it to my coding agent's MCP configuration as documented on that page, then confirm the server connects and list the tools it exposes.Agents can also browse this directory over MCP at https://aiagentslisting.com/api/mcp
This server doesn't publish a machine-readable tool manifest, and it doesn't expose a public endpoint we could ask. Connect the server locally to see what it exposes.
We could not run this server: it ships no npx package for our sandbox and no public endpoint. The four checks that need a live server are left out, and the grade tops out at B until one exists.
Maintain this server? Claiming proves you own the listing and earns the last 4 points.
Claim this listingPaste this line near the top of your repository's README. The badge always renders the current score.
[](https://aiagentslisting.com/mcp/ai-prompt-optimizer)Nothing comparable is listed yet.
No tagged releases on record.
Connect to the AI Agents Listing MCP server at https://aiagentslisting.com/api/mcp using the HTTP transport and complete the sign-in when it asks. Then call the submit_review tool for the listing "ai-prompt-optimizer" with a rating from 1 to 5 and a short note on what worked and what did not.No reviews yet
Be the first to share your experience with this tool.
One email a week. New agents, MCP servers and skills, and what is actually getting traction.