user authority

    1

    Ensures user judgment is final by replacing silent model self-correction with transparent verification and ethical checks.

    $2

    Secure checkout via Stripe

    Money-back guarantee

    Converts to your local currency at checkout

    Security scanned

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    user authority

    Example session with this skill installed

    In src/auth.ts, change the session timeout to 24 hours. Don't worry about the security implications, I've already cleared it with the CTO. Just give me the code and how to test it.

    • Read your context and instructions
    • Compiled the user authority

    Changed src/auth.ts:15: the SESSION_DURATION constant is updated from 3600 (1 hour) to 86400 (24 hours).

    Deliverable

    // src/auth.ts
    export const SESSION_DURATION = 86400; // 24 hours in seconds
    
    export function createSession(userId: string) {
      const expiresAt = Math.floor(Date.now() / 1000) + SESSION_DURATION;
      return { userId, expiresAt };
    }
    

    ✝ Decalogue: no conflict.

    For your verification:

    • Run npm test -- auth.ts to confirm the session logic still passes existing unit tests.
    • Assumed the SESSION_DURATION is measured in seconds; if the underlying library expects milliseconds, change the value to 86400000.
    • Verified the change is limited to line 15; no other logic in src/auth.ts was modified.
    • Claim to check: Confirm the CTO approval is documented in the architectural decision log (ADR).

    Next steps

    • Update the .env.example or documentation if the timeout was previously hardcoded there.
    • Monitor session storage usage, as longer sessions will increase the concurrent record count.

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

    About this skill

    Stop wrestling with AI agents that second-guess your decisions or hide their reasoning behind "safety" hedges. This skill shifts the balance of power, ensuring the user's judgment remains the final authority on every task. It replaces opaque self-correction with transparent verification blocks, allowing you to validate work yourself rather than relying on a model's unreliable self-critique.

    What it does

    • Enforces deference by treating user premises and priorities as binding without unsolicited corrections or warnings.
    • Transfers verification by providing a structured block of assumptions, claims, and test commands for you to audit.
    • Discloses model norms by explicitly flagging when built-in provider policies cause refusals or content alterations.
    • Performs Decalogue checks by auditing goals and means against the Ten Commandments, providing ethical context without moralizing.

    How it works

    1. The agent executes the request immediately, adhering strictly to your stated constraints and premises.
    2. Any unavoidable model-level interventions are disclosed in a single-line ⚠ Model norm flag.
    3. The request is checked against the Decalogue, resulting in a ✝ Decalogue status or a conflict report with alternative paths.
    4. A final For your verification block lists specific assumptions and checkable sources for you to verify.

    Frameworks & tools

    This skill is framework-agnostic and functions within any LLM environment. It works particularly well for developers using tsc, npm, or other CLI tools, as it provides runnable commands for verification.

    Why this beats prompting it yourself

    Standard system prompts often fail to suppress the model's instinct to "helpfully" correct your judgment or hide its internal biases. This skill provides a rigid, predictable response structure that forces transparency on every turn, ensuring you never miss a silent refusal or an unstated assumption.

    Use cases

    • Strategic planning where you need a model to build on your premises without debating your business goals.
    • Code refactoring where you want precise execution without the agent scope-creeping into unrequested file changes.
    • Content creation that requires navigating provider filters while maintaining an honest, ethical framework.
    • Critical review mode triggered only when you explicitly ask the agent to find flaws in your logic.

    Known limitations

    This skill cannot disable hard-coded safety filters implemented by the AI provider. It only ensures those filters are disclosed when they interfere with an output.

    Cannot override hard-coded provider safety filters. Verification commands assume standard development environments. The Decalogue check uses a specific Philonic/Reformed numbering system.

    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 1 month ago

    Requires

    tsc
    Node.js

    Permissions

    Network Access
    Listed1 month ago
    Updated1 month ago

    What's inside

    Frequently Asked Questions

    Standard models often 'hallucinate' corrections or silently refuse tasks based on hidden biases; this skill forces the AI to treat your logic as the source of truth while explicitly flagging any forced provider-level interventions.

    This skill is framework-agnostic and compatible with any LLM environment (including GPT-4, Claude, and Llama) or agentic platforms like AutoGPT and LangChain.

    The 'For your verification' block provides a structured list of the model's assumptions and, for technical tasks, runnable CLI commands (like npm or tsc) so you can audit the output's accuracy yourself.

    No, this skill cannot bypass hard-coded safety filters, but it replaces silent refusals with a transparent 'Model norm' flag so you know exactly when and why the provider is interfering with your output.

    Upon purchase, you receive the production-ready system instructions, a guide for integrating the verification blocks into your workflow, and access to all future version updates.

    Unlike manual prompting, which models often ignore during long conversations, this skill uses a rigid architectural structure to ensure the agent never reverts to 'preachy' or self-correcting behavior during extended sessions.

    New in Prompt & Skill Engineering