Fraud Risk Assessment: A Practical Agent Skill for Auditors

Where this fits

A financial auditor is preparing a year-end assessment for a client with revenue pressure, a new overseas subsidiary, and a whistleblower tip about procurement. Instead of starting from a blank template, they need a structured framework to evaluate fraud risk systematically. This skill packages the fraud triangle, ACFE fraud tree, red-flag checklists, and control evaluation matrices into a reusable capability their agent can apply consistently.

Why agents benefit

• Can classify a suspected scheme using the ACFE fraud tree — distinguishing skimming from billing schemes or expense-reimbursement fraud
• Applies the fraud triangle systematically, generating a risk narrative covering pressure, opportunity, and rationalization for each identified risk
• Runs red-flag checklists against transaction data, surfacing anomalies like Benford's-law deviations or unusual journal entries
• Produces structured output that maps directly to COSO/SOX documentation requirements
• Supports both periodic assessments and ad-hoc investigations triggered by whistleblower reports

Practical scenarios

Pre-M&A Fraud Screening

An agent reviews target-company financials, identifies related-party transactions and unusual revenue recognition, and flags high-pressure indicators like aggressive growth targets. It generates a risk memo for the deal team.

Annual COSO/SOX Compliance Assessment

The agent walks through the 17 COSO principles, applies the fraud-tree taxonomy across business units, and produces a control-deficiency matrix with recommended remediations.

Whistleblower-Triggered Investigation

Given a tip about procurement kickbacks, the agent identifies the relevant corruption scheme, maps it on the fraud tree, and creates an evidence-collection checklist.

Add it to your agent workflow

Install the skill:

npx skills add guoliang1114-boop/AriaAI --skill fraud-risk-assessment

Then instruct your agent with a prompt like:

Use the fraud-risk-assessment skill to evaluate this scenario: the sales
 director's compensation is tied to quarterly revenue, the finance team 
has no segregation of duties, and two suppliers share the same CEO. 
Produce a risk rating and recommended controls.

The agent will structure its analysis around the fraud triangle, classify each issue on the fraud tree, and return control recommendations.

Compare before adopting

Before adopting, assess: does the skill's fraud-tree coverage match your audit domain — for example, does it adequately cover corruption and financial-statement fraud, not just asset misappropriation? Check maintenance freshness — how recently was it updated and does it reflect current ACFE standards? Compare against alternatives like generic risk-assessment prompts or in-house templates, and decide whether the structured framework adds value over your existing workflow.

Why it is worth tracking

With 37 stars and a stable v1.0 status, this skill has enough community traction to warrant a closer look, especially if you build audit or compliance automation. The author continues to iterate on the AriaAI repo, so evaluate it again when new versions land or when you are scoping a broader anti-fraud program. Full details, updated docs, and examples are available at https://www.openagentskill.com/skills/guoliang1114-boop-fraud-risk-assessment.