Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    ap control testing and evidence collector

    1

    Manual testing of accounts payable controls is slow and prone to human error.

    $2.99

    Secure checkout via Stripe

    Money-back guarantee

    Converts to your local currency at checkout

    Security scanned
    ap control testing and evidence collector

    ap control testing and evidence collector

    Example session with this skill installed

    Run a control test on the attached April invoice extract using the $5k limit for Junior Approvers and $50k for Managers. Check three-way match with a 2% tolerance and 5-day exception SLA.

    • Read your context and instructions
    • Compiled the ap control testing
    • Generated the data export

    The AP control test for April has been completed using the provided parameters. A total of 150 invoices were analyzed against the approval matrix and three-way match rules.

    AP Control Testing Evidence Report — Acme Corp, 2024-05-15

    1. Scope, sources and assumptions

    • Authoriser: Internal Audit
    • Business Units: North America Operations
    • Sources: E1 (April_Invoices.csv), E2 (PO_Export.json)
    • Counts: 150 records processed; 0 out-of-scope.
    • Parameters: 2.00% match tolerance, 5-day exception SLA.

    2. Control results

    ControlTestedPassedFailedNot testableNot applicableDeviation rateConclusion
    Approval Limits1501428000.0533Deviations Found
    Segregation of Duties1501491000.0066Deviations Found
    Three-Way Match12011550300.0416Deviations Found
    Exception SLA54101450.2000Deviations Found

    3. Exception ledger

    InvoiceAmountControlResultReason codesEvidence id
    INV-8826500.00Approval LimitsFailLIMIT_EXCEEDEDEV-001
    INV-9011200.50Three-Way MatchFailPRICE_TOLERANCEEV-002
    INV-944450.00Segregation of DutiesFailSAME_USER_ENT_APPEV-003

    4. Evidence index

    • Approval Limits: 8 IDs (EV-001, EV-004..EV-010)
    • Segregation of Duties: 1 ID (EV-003)
    • Three-Way Match: 5 IDs (EV-002, EV-011..EV-014)

    5. Human-approval queue (drafts)

    Queue idInvoiceAmountReleasedFailed controlsProposed action
    Q-001INV-8826500.00NoApproval LimitsRoute to Senior Manager
    Q-002INV-944450.00YesSegregation of DutiesManagement decision required

    6. Limitations and warnings

    • Warning W-02: Missing receipt data for 30 invoices; Three-Way Match defaulted to PO-Invoice only.
    • Information gap: Organizational validation required for user roles assigned to INV-944.

    7. Observations and items for human review

    • Observed 5.33% deviation rate in approval limits.
    • Analytical estimate — validation required: Potential systemic delay in exception logging for BU-West.
    • One transaction (INV-944) violated segregation of duties and was already released.

    This testing supports, and does not replace, professional judgment and management's decision.

    Next steps

    • Review the 8 approval limit deviations in Section 3.
    • Verify user roles for the segregation of duties conflict on INV-944.
    • Provide receipt data for the 30 "Not applicable" match cases to re-run full three-way testing.

    April_Invoices.csv

    CSV · data export

    Generated

    Example file from a real run - the skill writes it into your workspace.

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

    About this skill

    Manual testing of accounts payable controls is slow and prone to human error. This skill automates the verification of approval limits, segregation of duties, and three-way matching using your raw data extracts.

    What it does

    • Approval verification checks invoice amounts against your specific approval matrix and individual user limits.
    • Segregation of duties identifies transactions where the same user performed conflicting actions like entering and approving.
    • Three-way matching validates consistency between purchase orders, receipts, and invoices within defined tolerances.
    • SLA monitoring tracks whether match exceptions were resolved within the required timeframe by independent parties.
    • Evidence generation produces a structured ledger and exception report for internal audit or management review.

    How it works

    1. Provide your data extracts including invoices, purchase orders, receipts, and your current approval matrix.
    2. The agent registers sources, validates inputs against a schema, and prepares a deterministic test request.
    3. A specialized script executes the tests to ensure no manual recomputation or rounding errors occur.
    4. The agent generates a structured report containing evidence IDs, deviation rates, and a draft human-approval queue.

    Frameworks & tools

    This skill uses a deterministic Python testing engine for calculations. It requires data in structured formats such as JSON or CSV extracts corresponding to the provided input schema.

    Why this beats prompting it yourself

    Generic prompts often hallucinate calculations or miss subtle segregation of duty conflicts. This skill uses a hardened script to ensure every calculation is traceable to an evidence ID, providing the precision required for financial controls.

    Use cases

    • Perform a month-end review of AP transactions to identify control deviations.
    • Collect evidence for internal auditors regarding exception resolution SLAs.
    • Verify that large vendor payments followed the correct approval hierarchy.
    • Audit segregation of duties across the enter, approve, and release lifecycle.

    Known limitations

    This is a read-only tool and cannot post to an ERP or release payments. It requires an authorizer and business unit scope to be defined before processing.

    Read-only only. Cannot approve or release payments. Requires specific fields like approval_matrix and authorized_by to run. No support for duplicate detection or fraud analysis.

    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 3 days ago

    Permissions

    Terminal / Shell
    Read Files
    Write Files
    Environment Variables
    Listed3 days ago

    What's inside

    Frequently Asked Questions

    This skill automates the verification of approval limits, three-way matching, and segregation of duties. It replaces manual sampling with a deterministic testing engine that eliminates human error and provides 100% coverage of your transactional data.

    The skill is designed to work with structured data extracts (CSV or JSON) from any ERP system. As long as your data can be mapped to the required schema for invoices, POs, and receipts, the agent can process it regardless of your underlying accounting software.

    No, this is a read-only auditing tool. It is designed strictly for testing and evidence collection, meaning it cannot post entries, change approval limits, or release payments in your production environment.

    You will receive the AI agent skill configuration, the deterministic Python testing engine, and the standardized input schemas required to format your data for successful testing.

    Unlike generic LLM prompts that may hallucinate calculations, this skill uses a hardened, script-based testing engine. This ensures that every exception is traceable to a specific evidence ID and that all math is 100% accurate for financial reporting.

    Setup involves mapping your current approval matrix and business unit scope to the tool's input template. Once your data fields are aligned with the schema, the agent can begin running tests and generating evidence immediately.