Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    ai agent governance audit

    2

    Audit agentic systems against runtime controls and generate compliance artifacts for ISO 42001 and NIST AI RMF.

    $16

    Secure checkout via Stripe

    Money-back guarantee

    Converts to your local currency at checkout

    Security scanned
    ai agent governance audit

    ai agent governance audit

    Example session with this skill installed

    We have a customer-support agent that can issue refunds up to $50. It uses a system prompt to stay in scope and a service account for Shopify. Is this safe to ship?

    • Read your context and instructions
    • Compiled the ai agent governance

    Finding: FAIL. System prompts are requests, not enforcement; prompt injection can bypass refund limits. Finding: FAIL. Service account lacks least-privilege. Status: UNVERIFIED for Containment. Recommendation: Implement a runtime cap at the API level and drill the kill switch before deployment.

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

    About this skill

    The problem

    Most AI governance is just a policy binder that ignores how agents actually behave. When an agent holds API keys, makes purchases, or deletes data, a "system prompt" is not a security control, and an unproven kill switch is a liability.

    What it does

    • Audits deployed agents against seven critical runtime controls, including containment, permission scope, and blast radius.
    • Differentiates between PASS, FAIL, and UNVERIFIED status based on evidence rather than architectural claims.
    • Identifies "decorative" human oversight where approval fatigue renders the human-in-the-loop ineffective.
    • Produces audit-ready findings mapped to ISO 42001, NIST AI RMF, and OWASP Agentic AI Top Ten.
    • Generates interview scripts and demonstration lists to force teams to prove their safety claims.

    Why this beats prompting it yourself

    General LLMs treat system prompts as enforcement and often hallucinate framework control IDs. This skill enforces a "test, don't trust" logic that prioritizes exploitability and runtime reality over paperwork, ensuring you don't submit fabricated citations to a real auditor.

    Use cases

    • Pre-deployment security reviews for agents with write-access to production databases or financial tools.
    • Post-incident reconstruction to determine how and why an agent performed an unintended action.
    • Preparing evidence and artifacts for ISO 42001 certification or NIST framework compliance.
    • Validating that multi-agent systems have sufficient circuit breakers to prevent cascading failures.

    Known limitations

    This is not a certification or attestation. It does not cover ML model bias or fairness auditing. It requires manual input of architectural details or interview responses to function.

    Not for ML bias auditing. Produces audit-prep findings, not legal certifications. Requires user to provide architectural details or test results.

    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 4 months ago

    Permissions

    No special permissions declared or detected

    Listed4 months ago

    What's inside

    Frequently Asked Questions

    It tests runtime controls against the actual agent — can it be killed, what can it reach, can last week's action be reconstructed — rather than checking whether a policy document exists. Framework mapping (ISO 42001, NIST AI RMF) is included where the framework text was actually read and cited; anything unread is marked "TO VERIFY" rather than guessed. Findings are audit-prep and control evidence, not a certification or attestation.

    This audit skill is framework-agnostic and can assess any agentic system, provided you can supply details on its permissions, API access, and operational guardrails.

    You receive the seven-control audit method (containment, permission scope, purpose limitation, audit trail, human oversight, blast radius, evaluation), interview-script and demonstration-list templates for testing each control, and framework mapping to ISO 42001 / NIST AI RMF / OWASP Agentic AI Top Ten — cited from the actual framework text where read, marked "TO VERIFY" where not. It's a procedure an agent runs against your system, not an automated crosswalk generator.

    No, this skill focuses on security architecture and runtime enforcement; it does not measure algorithmic bias, data toxicity, or dataset fairness.

    Setup is immediate upon purchase; you simply provide the agent's architectural details or runtime logs to begin generating audit-ready compliance evidence.

    The skill is frequently updated to reflect the latest versions of the OWASP Agentic AI Top Ten and evolving NIST AI RMF requirements to ensure your audits remain current.

    New in Agents & Orchestration