Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIDeepSeekDeepSeekMuseMuse+22 more

    AI Governance Audit-Evidence Reviewer

    1

    Map AI governance controls to supplied evidence and identify stale, contradictory or missing support without certifying compliance.

    $16.99

    Secure checkout via Stripe

    Money-back guarantee

    Converts to your local currency at checkout

    Security scanned
    AI Governance Audit-Evidence Reviewer

    AI Governance Audit-Evidence Reviewer

    Example session with this skill installed

    Review the supplied AI system inventory, governance policy, model evaluation report and approval records. Map human-oversight and monitoring controls to evidence, identify gaps or contradictions, and return a traceable exception register with passed, skipped and suppressed checks.

    • Read your context and instructions
    • Compiled the ai governance audit-evidence

    Review summary

    Fourteen eligible control checks were performed across the governance policy, system inventory and evaluation records. Ten checks passed, two evidence-backed review points were identified, and two checks were skipped because the cited approval log and monitoring appendix were not supplied in usable form. One name-similarity candidate was suppressed after the system IDs were shown to differ.

    Exception register

    IDClassificationControl/locationFindingEvidenceRecommendationSeverityConfidence
    AIGR-001Confirmed exceptionMonitoring control GOV-02; evaluation report page 4The recorded bias evaluation is older than the frequency required by the supplied policy.Last test 142 days before review; policy maximum 90 days.Perform or evidence the required review and document accountable approval.HighHigh
    AIGR-002Review promptInventory row 22; risk approval registerThe system is classified as low risk, but the supplied register does not show the required approval.System ID CHAT-22; approval field blank; policy section 5.3 requires approval.Confirm whether the approval exists outside the supplied pack.MediumHigh

    Checks passed

    • Inventory owner and version fields agreed to the evaluation report for tested systems.
    • Human-oversight evidence was present for three in-scope high-risk systems.
    • Incident status and closure dates were internally consistent.

    Checks skipped

    • Approval-authority verification: the referenced approval log was not supplied.
    • Drift-threshold verification: the monitoring appendix was truncated.

    Suppressed candidate

    • Two similarly named assistants had different system IDs and owners; no duplicate-inventory finding was raised.

    This review is limited to supplied evidence and does not certify regulatory or policy compliance. Final conclusions remain with accountable management and the responsible professional.

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

    About this skill

    Map AI governance controls to supplied evidence and identify gaps, contradictions and stale records through a disciplined review workflow.

    Best for: Internal audit, risk, compliance, model-risk and governance teams preparing AI control evidence for management or assurance review.

    The skill distinguishes a policy statement from evidence that a control operated. It maps criteria to records, tests dates and consistency, and never treats an omitted file as proof of control failure.

    What you receive

    • control-to-evidence matrix with exact page, section and record locations
    • evidence-backed exception and review-prompt register
    • stale, contradictory and missing-evidence analysis
    • separate passed, skipped and suppressed-check schedules

    What it reviews

    • system inventories and risk classifications
    • approvals, human oversight and accountability records
    • model evaluation, monitoring and incident evidence
    • vendor and third-party assessment consistency
    • policy requirements versus operating evidence
    • dates, owners, versions and review frequency

    How false positives are controlled

    Missing evidence is classified as a skipped or incomplete-evidence check unless the supplied criteria require the record to exist in scope. Policy wording alone is not treated as proof of operation, and ambiguous OCR is not elevated without corroboration.

    Review boundary

    This is an evidence-review aid. It does not certify compliance, issue an audit opinion, provide legal advice or replace accountable management and professional judgement.

    Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Business & Operations.

    No legal/regulatory certifications. Cannot infer inventory completeness from partial exports. Requires user-supplied policy criteria. High-confidence findings require clear source record IDs.

    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

    Money-back guarantee

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

    Trust & safety

    Security scan

    Security scan passed 14 days ago

    Permissions

    Read Files
    Write Files

    Read supplied governance, policy and evidence files. Write access is used only when the user requests an exported control matrix or review report. No network access or credentials required.

    Listed14 days ago

    What's inside

    Frequently Asked Questions

    Unlike generic LLMs that might assume evidence exists, this skill uses a deterministic review process that requires exact record IDs and page references, preventing the 'hallucination' of compliance where documentation is actually missing.

    The skill is compatible with major AI agents and GRC platforms that support document ingestion; it works by analyzing your uploaded model inventories, evaluation reports, and policy frameworks against your specific control criteria.

    No, this skill is a diagnostic tool that flags gaps and inconsistencies; it provides a structured exception report for human auditors to review rather than issuing a final certification or legal opinion.

    You receive the full evidence review workflow, including pre-configured mapping logic, arithmetic reperformance modules, and a structured exception reporting template ready for integration into your audit trail.

    The skill is updated periodically to reflect emerging AI governance standards and to improve its extraction capabilities for new technical report formats commonly used in AI evaluation.

    Setup involves defining your scope mapping and uploading your policy framework; once your inventory of files is established, the skill performs a preliminary 'Extraction Assessment' to ensure all data is readable before the full audit begins.

    New in Business & Operations