Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    bix-claude-skills: Prompt Rewriter With Frameworks

    0

    Transform raw prompts into high-performance instructions using 20+ specialized prompt engineering frameworks.

    Free

    1 useSecurity scanned
    bix-claude-skills: Prompt Rewriter With Frameworks

    bix-claude-skills: Prompt Rewriter With Frameworks

    Example session with this skill installed

    /rewrite "Write a script that renames all the photos in a folder by date"
    

    Or in plain language: "Rewrite this prompt for me with a good framework: write a script that renames all the photos in a folder by date."

    • Read your context and instructions
    • Compiled the bix-claude-

    Summary: the skill read the prompt and the session context (a Python utility project, Python 3.11, no third-party packages). The recommender suggested TIDD-EC and offered Chain of Thought as the runner-up. You accepted the recommendation, and the rewriter produced the prompt below. Items tagged [inferred] were not in your original text, so check them.

    Question shown to you

    Based on your prompt, I recommend TIDD-EC. How would you like to proceed?
    1.

    Use TIDD-EC (recommended): The prompt asks for a script with specific naming behavior, which suits precise do/don't rules.

    2.

    Use Chain of Thought instead: Better if you want the model to reason through edge cases such as missing dates before coding.

    3. Browse all frameworks
    4. Custom / Skip

    Rewritten using TIDD-EC

    Task
    Write a Python script that renames every photo in a folder based on its date.
    
    Instructions
    Accept a folder path as a command-line argument. For each image file in the folder, read the date the photo was taken and rename the file to YYYY-MM-DD_HHMMSS.ext. Print each old name and new name. [inferred] Use the file modification time when no capture date is stored in the file.
    
    Do
    - Support common photo extensions (.jpg, .jpeg, .png). [inferred]
    - Handle two photos with the same timestamp by adding a numeric suffix.
    - Offer a dry-run flag that prints changes without renaming. [inferred]
    
    Don't
    - Do not overwrite existing files.
    - Do not rename files that are not images.
    - Do not recurse into subfolders. [inferred]
    
    Examples
    IMG_4021.jpg (taken 2024-06-03 14:22:09) becomes 2024-06-03_142209.jpg. [inferred]
    
    Context
    Part of a Python 3.11 utility project that uses only the standard library.
    

    Then you were asked: Use this prompt / Try another framework / Edit first. You chose "Use this prompt", and the skill output:

    Now executing the rewritten prompt

    (the prompt above, verbatim, which the main session then acts on)

    Next steps

    • Check each [inferred] line and edit anything that is wrong.
    • Choose "Try another framework" to compare, for example, Plan-and-Solve.
    • Choose "Edit first" to request changes before the prompt is run.

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

    About this skill

    Raw, conversational prompts often lead to vague or hallucinated outputs. This skill forces structure onto your requests by applying one of 20 proven prompt engineering frameworks, ensuring the LLM understands your intent, constraints, and required format.

    What it does

    • Framework Recommendation analyzes your prompt and suggests the best structure, such as RISEN for general tasks or TIDD-EC for precision coding.
    • Context Preservation summarizes your current session files and goals to ensure the rewrite remains relevant to your specific project.
    • Sub-agent Delegation offloads the heavy lifting to a fast Haiku sub-agent, keeping your main conversation history clean and focused.
    • Template Library provides instant access to specialized reasoning frameworks like Tree-of-Thought, Step-Back, and Chain-of-Density.
    • Custom Overrides allows you to supply your own .md or .txt templates if the built-in catalog does not fit your needs.

    How it works

    1. Input your raw prompt or have the skill capture the current context from your workspace.
    2. Select a recommended framework or browse the catalog by category (Reasoning, Creative, Expert, etc.).
    3. Approve the generated rewrite and choose to execute it immediately or refine the output further.

    Frameworks & tools

    This skill utilizes a library of 20 markdown-based prompt templates. It leverages sub-agent orchestration to process rewrites without polluting the main token window.

    Why this beats prompting it yourself

    Manually formatting complex prompts like Chain-of-Thought or CO-STAR is time-consuming and prone to missing structural tags. This tool automates the boilerplate, letting you focus on the core logic while ensuring the LLM receives perfectly structured instructions.

    Use cases

    • Refactoring vague feature requests into actionable coding specifications.
    • Applying Chain-of-Thought reasoning to complex debugging sessions.
    • Generating high-density documentation summaries using specialized frameworks.
    • Creating expert personas for code reviews or creative writing tasks.

    Known limitations

    The skill requires a sub-agent compatible environment. Rewriting performance depends on the clarity of the original input.

    • Built for Claude Code. It needs the Agent tool and AskUserQuestion, so it is not tested elsewhere.
    • Both the recommendation and the rewrite require sub-agents. If a sub-agent errors or returns nothing, the skill tells you and offers another framework.
    • If the recommender's reply cannot be parsed or names an unknown framework, the skill falls back to browsing the catalog.
    • The catalog is fixed at 20 frameworks. Anything else needs a custom template.
    • The conversation summary is capped at about 150 words, so long sessions are only roughly captured.
    • Rewrites restructure your prompt. They do not add goals, and guessed details are tagged [inferred] for you to verify.
    • Quality depends on the Haiku model used for rewriting. Review the result before using it.
    • "Skip rewriting" echoes the original prompt and ends the skill.

    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 any payment.

    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

    1 use

    Times used to date

    Free forever

    No account required to browse

    Trust & safety

    Security scan

    Security scan passed 1 month ago

    • Free with an Agensi account

    Permissions

    Terminal / Shell
    Read Files

    File Scopes

    ${CLAUDE_SKILL_DIR}/templates/*.md
    ${CLAUDE_SKILL_DIR}/references/agents/*.md

    The skill only reads its own template files and, if you choose, one template file you name. It sends your prompt and a short conversation summary to sub-agents inside your Claude Code session. It does not edit files, run project commands or contact outside services. The rewritten prompt is only run after you approve it.

    Listed1 month ago

    What's inside

    Frequently Asked Questions

    This skill automates the complex task of prompt engineering by instantly applying structured frameworks like RISEN or Chain-of-Thought to your raw requests, eliminating vague outputs and hallucinations.

    The skill includes a library of 20+ specialized frameworks, a sub-agent orchestration layer for clean processing, and the ability to import your own custom .md or .txt templates.

    Yes, this skill is designed to work in environments that support sub-agent delegation (such as Anthropic Haiku) to ensure your main conversation history remains focused and uncluttered.

    Absolutely—while the skill recommends the best framework for your task, you can manually browse the catalog by category or provide your own custom templates for full control.

    Setup is instantaneous upon installation; the skill automatically scans your workspace context to help you refactor prompts without any manual configuration or coding required.

    Yes. Choose Custom / Skip, then give a path to a .md or .txt file, or paste the template text.

    New in Prompt & Skill Engineering