Fraud Red-Flag Triage Assistant
Triage suspicious transactions and audit exceptions into evidence-linked briefs with alternative-explanation testing.
$6.99
Works with the AI Agents you already use
Claude Code
Cursor
Codex CLI
DeepSeek
Muse+22 moreClaude CodeCursorCodex CLIDeepSeekMuse
Dots
Manus
Grok
Grok Bot+18 agentsFraud Red-Flag Triage Assistant
Example session with this skill installed
Review two invoices from Vendor V-102 for $4,800 each, submitted on consecutive days. Our single-approval threshold is $5,000. Provide a triage brief.
- Read your context and instructions
- Compiled the fraud red-flag triage
I have analyzed the submitted invoice data against the $5,000 approval threshold. The following triage brief identifies potential indicator patterns and required validation steps.
Fraud Red-Flag Triage Brief
1. Scope, status and sources
Purpose: Initial triage of potential invoice splitting.
Sources: User-provided invoice amounts and dates.
Evidence Quality: Preliminary; internal policy not yet formally verified against system logs.
2. Fact and evidence register
| Fact/observation | Source ID | What it supports | Confidence | Validation |
|---|---|---|---|---|
| Inv #1: $4,800 (Date T) | User Input | Near-threshold payment | High | Verify Invoice PDF |
| Inv #2: $4,800 (Date T+1) | User Input | Near-threshold payment | High | Verify Invoice PDF |
| Total: $9,600 | Calculated | Exceeds $5k threshold | High | Confirm Policy Doc |
3. Indicator and alternative-explanation matrix
| Indicator | Supporting evidence | Non-fraud explanation | Test |
|---|---|---|---|
| Potential Split Invoicing | Two payments to same vendor < 24h apart totaling > limit | Vendor issued separate bills for distinct deliverables | Review Purchase Order line items |
| Approval Bypass | Combined total exceeds manager limit | Emergency procurement authorized by higher spend limit | Check delegation of authority log |
4. Preliminary triage category
Category: Medium
Rationale: The proximity of the transactions to the $5,000 threshold and each other is a classic indicator of limit circumvention, but missing PO data prevents a "High" classification.
5. Evidence and preservation request
| Request | Purpose | Custodian | Priority |
|---|---|---|---|
| Purchase Orders | Check if deliverables were split | Procurement | High |
| Approval Log | Identify who authorized the payments | Finance Ops | High |
Next steps
- Upload the specific Purchase Orders associated with Vendor V-102.
- Confirm the exact date the $5,000 threshold policy was implemented.
- Share the Triage Brief with the Internal Audit lead for authorization of further review.
Connects securely to your tools. The creator never sees your data.
About this skill
Manual fraud triage is slow, biased, and prone to overlooking benign explanations. This skill structures raw transaction data and allegations into a reproducible, evidence-linked triage report for audit and compliance teams.
What it does
- Indicator mapping: Identifies specific anomalies like split invoices, duplicate payments, or unauthorized beneficiary changes based on facts.
- Alternative testing: Proposes non-fraud explanations such as clerical errors or batch processing to prevent premature conclusions.
- Chronology building: Generates a normalized timeline of events, tracing initiators, approvers, and payment routes.
- Evidence sequencing: Drafts prioritized requests for bank references, access logs, and contracts with privacy safeguards.
- Escalation briefing: Produces a neutral, confidential brief for authorized management review.
How it works
- Submit transaction records, case notes, or suspected control exceptions.
- The agent inventories sources, normalizes timestamps, and builds a fact register.
- It maps indicators against evidence while simultaneously seeking counter-evidence or benign explanations.
- It outputs a structured Triage Brief including a prioritized action tracker and evidence preservation requests.
Frameworks & tools
This skill is platform-agnostic and works with raw text, CSV exports from ERP systems like SAP or Oracle, and internal audit logs.
Why this beats prompting it yourself
Generic prompts often jump to conclusions or fail to account for legitimate business exceptions. This skill enforces a strict methodology that separates facts from inferences and ensures every "red flag" is linked to a specific source ID and a testable alternative explanation.
Use cases
- Reviewing procurement anomalies for potential vendor-manager collusion.
- Triaging internal tips regarding unusual expense claims or payroll exceptions.
- Auditing existing fraud reports to identify overstated claims or insufficient evidence.
- Organizing evidence for confidential escalation to legal or HR departments.
Known limitations
Does not determine guilt or legal liability. It requires structured or semi-structured transaction data to be effective and cannot verify the physical authenticity of uploaded documents.
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.
Cannot establish legal intent or guilt. Requires manual input of organizational policies/thresholds if they are to be used for comparison. Does not perform automated data extraction from images.
How to use
- 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
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 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 9 days ago