Prompt Optimizer — Production Prompt Engineering

    1

    Turn business requirements into concise, actionable prompts for AI products. Built around token efficiency, response speed, and output consistency.

    Free

    Security scanned

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIDeepSeekDeepSeekMuseMuse+22 more

    Prompt Optimizer — Production Prompt Engineering

    Example session with this skill installed

    Synthetic example from the bundled reference cases.

    Use Prompt Optimizer to simplify the prompt below. Keep the same selection criteria, JSON fields, missing-value behavior and {{meeting_notes}} placeholder. Show the complete revised prompt.

    ORIGINAL PROMPT
    你负责从会议记录中整理待办,需要认真理解每个人的发言,准确识别真正需要完成的工作。
    处理时先完整阅读会议记录,再筛选待办。要区分明确承诺、一般建议和已经完成的事情,只留下参会人明确承诺且尚未完成的事项。没有承诺的建议不算待办,已经完成的事项也不要留下。
    对每条待办识别任务、负责人和期限。负责人或期限没有明确说出时不要猜测,对应字段填 null。没有符合条件的事项时返回空数组。
    最终只输出 JSON 数组,每项有 task、owner、deadline 三个字段。输出前再次检查每条事项都属于明确承诺且尚未完成的工作,不要加入解释。
    会议记录:
    {{meeting_notes}}

    • Read your context and instructions
    • Compiled the prompt optimizer

    Revised prompt

    从会议记录中提取参会人明确承诺且尚未完成的待办。只输出 JSON 数组,每项包含 task、owner、deadline;负责人或期限未明确时填 null,没有符合条件的事项时返回 []。
    
    会议记录:
    {{meeting_notes}}
    

    Key changes

    Combined repeated selection rules into one explicit condition. Preserved the JSON contract, null values, empty-result branch and input variable. No workflow headings were added because this task does not need them.

    Synthetic check

    Input

    林晓:我周五之前把接口文档补好。
    陈宇:也许可以加个筛选按钮。
    林晓:上周的故障复盘已经发出了。
    周岚:我来核对那份清单,时间还没定。
    

    Expected output from the revised prompt

    [{"task":"补全接口文档","owner":"林晓","deadline":"周五之前"},{"task":"核对清单","owner":"周岚","deadline":null}]
    

    This is a semantic illustration, not a live model benchmark.

    Connects securely to your tools. The creator never sees your data.

    About this skill

    Prompt Optimizer helps developers, founders, and domain experts improve the prompts behind their AI products—without having to learn prompt engineering first.

    Give your agent a requirement or an existing prompt. It clarifies business rules, removes instructions that add no value, and organizes complete examples where they help. You get a revised prompt that is easier to understand, test, and maintain.

    The approach comes from hands-on prompt refinement in AI projects, with three priorities:

    • Cost and speed: reduce unnecessary instructions while preserving the information the model needs.
    • Output consistency: make conditions, decision order, and expected results explicit.
    • Maintainability: keep requirements easy to locate and update as the product evolves.

    Useful for conversational AI, knowledge Q&A, content generation, and information extraction—including prompts intended for domain-specific and on-device models. ICTIO provides a flexible structure; simple tasks stay simple.

    How to use

    Connect your agent to Agensi, sign in, and ask it to use Prompt Optimizer. Paste this request followed by your requirements or existing prompt:

    Use Prompt Optimizer to improve the prompt below. Preserve my business requirements, variables, and output format. Ask about ambiguities that affect the result, then give me the complete revised prompt and briefly explain the changes.

    Release notes

    Instructions and bundled examples are primarily in Chinese. This is an initial release, improved through concrete feedback and failure cases. Check revisions against your own cases and target model; the bundled examples are illustrations, not benchmark results.

    Source, documentation and feedback: https://github.com/EchoUser005/prompt-optimizer

    The included source is MIT-licensed; the package retains its copyright and license notice.

    Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Prompt & Skill Engineering.

    Primarily Chinese instructions and examples. Produces a reviewable candidate, not guaranteed performance gains. Does not deploy prompts, invoke paid models, or integrate with production systems automatically. Evaluation guidance is included; users provide cases, model access and results when needed.

    How to use

    1. 1

      Connect your agent to the Agensi MCP

      Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup

    2. 2

      Sign in to Agensi

      Your agent shows a sign-in link. Sign in and click Allow.

    3. 3

      Ask for what you need

      Your agent loads this skill and uses it on your task. It always asks before buying a skill.

    Reviews

    No reviews yet

    Be one of the first to try it. Every listed skill goes through the trust checks below.

    Security scanned

    Checked by our 8-point scan before listing

    Fresh listing

    Recently published to Agensi

    Free to use

    No account required to browse

    Trust & safety

    Security scan

    Security scan passed 15 days ago

    • Free with an Agensi account

    Permissions

    No special permissions declared or detected

    Listed15 days ago

    What's inside

    Frequently Asked Questions

    The skill itself is text-only and requires no extra API key or service. Your agent and any live evaluation model have their own requirements.

    No. ICTIO separates content by responsibility. Only use the sections that the task needs; a simple task may remain a single paragraph.

    No cross-model performance claim is made. The package includes synthetic examples and guidance for controlled comparisons using your own cases.

    New in Prompt & Skill Engineering

    Free