Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    4d analyzer

    1

    Audit and rewrite prompts or system files using Anthropic's 4D AI Fluency framework.

    $5

    Secure checkout via Stripe

    Money-back guarantee

    Converts to your local currency at checkout

    Security scanned
    4d analyzer

    4d analyzer

    Example session with this skill installed

    Review this prompt for a technical blog post generator: "Write a 500 word blog post about Kubernetes networking for a junior dev audience. Focus on Services and Ingress."

    • Read your context and instructions
    • Compiled the 4d analyzer

    I have analyzed your prompt against the 4D AI Fluency framework. You are operating in Automation mode.

    4D Fluency Report: Kubernetes Blog Post

    DimensionVerdictEvidence / Feedback
    DelegationStrongThe task is well-suited for an LLM. You have clearly defined the scope (Services and Ingress).
    DescriptionAdequateAudience Description is present ("junior dev"), but Process Description is missing. No style or structure constraints.
    DiscernmentNot AssessableRequires the AI's output to evaluate how you judge the result.
    DiligenceNot AssessableRequires the full conversation history to see how you verified the technical accuracy.

    Framework Analysis

    Missing Context: You specified the "what" but not the "how." Without a defined structure (e.g., "compare and contrast"), the AI may produce a generic list.

    Process Gap: There is no instruction on how to handle technical jargon, which is critical for a "junior dev" audience.

    Rewritten Prompt

    Write a 500-word technical blog post explaining Kubernetes networking, specifically focusing on the relationship between Services and Ingress.

    Audience: Junior DevOps engineers who understand containers but are new to cluster networking.

    Structure: Start with the "why" (problem), explain Services (internal), then Ingress (external), and end with a simple analogy.

    Tone: Educational and encouraging. Define technical terms like "ClusterIP" and "NodePort" when first mentioned.
    Constraint: Avoid using complex YAML examples; focus on the conceptual flow of traffic.

    Next steps

    1. Run the rewritten prompt and share the output to assess Discernment.
    2. Check if the AI correctly simplified the "ClusterIP" definition for the junior audience.

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

    About this skill

    Most prompt engineering advice relies on generic tips like "give it a persona." This skill applies the structured Anthropic AI Fluency 4D framework—Delegation, Description, Discernment, and Diligence—to evaluate how humans and AI actually collaborate. It moves beyond "good prompting" to measure systemic competence in AI interaction.

    What it does

    • Multi-dimensional analysis evaluates inputs against specific sub-competencies like Process Description and Output Discernment.
    • Evidence-based grading assigns Strong, Adequate, Weak, or Not Assessable status based strictly on the provided text.
    • Context-aware assessment adapts its critique based on whether the interaction is Automation, Augmentation, or Agency.
    • Framework-anchored feedback identifies specific failures, such as missing audience context or inappropriate task delegation.
    • Structural rewriting provides a version of the prompt or system file optimized for higher fluency scores.

    How it works

    1. Input Classification determines if you provided a single prompt, a full conversation, or a system prompt/SKILL.md file.
    2. Mode Identification labels the interaction style to set the benchmark for what "good" looks like for that specific task.
    3. Assessment runs the input against the 4D evidence matrix, quoting specific passages to justify every verdict.
    4. Correction generates an improved version of the input that addresses identified gaps without adding unnecessary "prompt fluff."

    Frameworks & tools

    This skill is built on the Anthropic AI Fluency Framework by Dakan and Feller. It uses a structured reporting format designed for technical environments like Claude Code or Cursor.

    Why this beats prompting it yourself

    Generic LLMs often give "polite" feedback that pads prompts with useless ceremony. This skill uses a rigid evidence matrix to ensure feedback is anchored in research-backed competencies and refuses to "hallucinate" grades for dimensions it cannot see.

    Use cases

    • Prompt Auditing to fix prompts that consistently deliver off-target or low-quality results.
    • Skill Optimization for developers building agentic workflows who need to harden their SKILL.md files.
    • Conversation Review to understand where a multi-turn interaction went off the rails.
    • Delegation Check to determine if a task is actually suitable for an LLM or if it needs to be broken down.

    Known limitations

    The skill can only assess competencies visible in the input; for example, it cannot judge Discernment from a single prompt without its output. It will not improve prompts designed to cause harm.

    Cannot assess Discernment or Diligence from a single prompt without its output. Requires explicit input classification to provide accurate grades. Will not optimize harmful requests.

    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

    Fresh listing

    Recently published to Agensi

    Money-back guarantee

    Doesn't do what its page promises? You get your money back.

    Trust & safety

    Security scan

    Security scan passed 28 days ago

    Permissions

    Read Files
    Write Files

    4d-analyzer is fully self-contained. I grepped its SKILL.md and all references/ files for any call-out to another skill (by name, Skill(...) invocation, or /plugin install reference), and found nothing. The only "skill" match was the word "skill" used generically in evidence-matrix.md to refer to itself. The 4D framework definitions, the evidence matrix, the rewriting guidance, the report template and it vendored locally in its own references/assets/.

    Listed28 days ago
    Updated7 days ago

    What's inside

    Frequently Asked Questions

    The skill audits your inputs against the four dimensions of Delegation, Description, Discernment, and Diligence to identify specific structural weaknesses that cause generic or off-target AI responses.

    This skill is optimized for structured environments like Claude Code and Cursor, and it is compatible with any agentic workflow that utilizes system files or SKILL.md documentation.

    Yes, the analyzer identifies the interaction mode—Automation, Augmentation, or Agency—and adjusts its grading criteria to reflect the specific competence required for that level of collaboration.

    You receive the 4D Analyzer skill logic, the evidence-based grading matrix, and documentation on how to integrate the analyzer into your existing development or prompt engineering workflow.

    Updates are provided periodically to reflect new research in AI fluency and framework refinements, ensuring your assessments remain aligned with the latest industry benchmarks.

    New in Prompt & Skill Engineering