Works with the AI tools you already use
Claude Code
Cursor
Codex CLI
Muse
OpenClaw+21 moreClaude CodeCursorCodex CLIMuseOpenClaw
Manus
Grok
Grok Bot+18 agentsPhotoAudit: InsurTech Document Completeness & Fraud Guard
Multimodal computer-vision guardrail for insurance claim compliance and automated fraud detection.
$149
PhotoAudit: InsurTech Document Completeness & Fraud Guard
Example session with this skill installed
Audit these two photos from a reported rear-end collision at a red light. The driver says they were hit from behind at 10mph. Images show a deep puncture in the trunk but no license plate for Car B.
- Read your context and instructions
- Compiled the photoaudit
- Generated the document
The evidence package has been audited for compliance and consistency. The submission fails the completeness threshold and shows significant geometric anomalies.
Audit Report: Case ID_99281-AUDIT
| Field | Value |
|---|---|
| Status | REJECTED |
| Documentation Gaps | ERR_NO_LICENSE_PLATE: Vehicle B plate not visible in frames; WARN_NO_LANDMARKS: Missing geospatial infrastructure markers. |
| Fraud Risk Flags | ERR_DAMAGE_MISMATCH: Impact described as 10mph rear-end strike, but image img_02 shows a deep point puncture inconsistent with a flat bumper strike. |
| Verdict Reason | Evidence package fails mandatory license plate visibility and shows damage geometry that contradicts the textual impact vector. |
{
"status": "REJECTED",
"verdict_reason": "Evidence package fails completeness and consistency thresholds. Routed to manual review.",
"audit_trail": {
"block_hash": "sha256:7f83b1a23c4d5e6f7g8h9i0j1k2l3m4n5o6p7q8r9s0t1u2v3w4x5y6z7a8b9c0d",
"timestamp": "2024-05-20T14:22:15Z",
"verdict": "REJECTED"
}
}
Next steps
- Trigger an automated request to the driver for a photo showing Vehicle B license plate.
- Escalate Case ID_99281 to the Special Investigations Unit (SIU) for manual damage depth verification.
- Log the
ERR_DAMAGE_MISMATCHflag to the claimant's risk profile.
photoaudit-insurtech-document-completene.pdf
PDF · document
Example file from a real run - the skill writes it into your workspace.
Connects securely to your tools. The creator never sees your data.
What you get
About this skill
Manual review of accident scene photos is slow, prone to human error, and expensive. This skill acts as a multimodal vision guardrail that automates the audit of insurance claim evidence for regulatory compliance and fraud indicators.
What it does
- Compliance auditing identifies missing license plates or required geospatial landmarks in accident photos.
- Trajectory validation checks wheel vectors and debris placement against the described impact.
- Damage geometry analysis detects physical inconsistencies between described low-velocity strikes and deep puncture damage.
- Fraud risk tagging flags evidence packages that show signs of staged accidents or manipulated scene telemetry.
- Audit trail generation produces immutable JSON records with block hashes and timestamps for regulatory defensibility.
How it works
- Input ingestion receives raw image assets and driver-submitted textual incident telemetry.
- Vision processing executes multi-object detection to locate plates, infrastructure, and vehicle contact profiles.
- Consistency mapping runs geometric cross-checks between physical evidence and the impact vector description.
- Verdict delivery returns a structured JSON report with a categorical
APPROVEDorREJECTEDstatus.
Frameworks & tools
Requires a multimodal model such as GPT-4V, Claude 3.5 Sonnet (Vision), or Gemini 1.5 Pro. It is designed for integration into Python or Node.js claims-processing backends via structured JSON payloads.
Why this beats prompting it yourself
Prompting vision models for insurance often results in vague descriptions or hallucinated confidence scores. This skill enforces a strict API contract with specific error codes like ERR_DAMAGE_MISMATCH, ensuring the output is immediately actionable by downstream automated systems.
Use cases
- Automating the initial intake gate for insurance mobile apps to reject incomplete photos.
- Adding a fraud detection layer to digital Europrotocol filing platforms.
- Pre-screening high volumes of claims to flag anomalies for manual investigative units.
Known limitations
Requires high-resolution imagery and direct access to incident telemetry. It does not provide legal advice or speculative fraud probabilities.
How to install
Works the same in every agent - Claude, Cursor, Codex, Copilot and 20+ more.
- 1
Download the ZIP
Free skills download straight away. Paid skills unlock right after purchase.
- 2
Unzip into your skills folder
Every agent reads skills from one folder on your machine. Drop the unzipped folder in there.
- 3
Ask your agent to use it
Restart the agent if it was already running. It picks the skill up automatically - no config needed.
Skills folder by agent
Click the path to copy it. Create the folder if it does not exist yet.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill passes our trust checks below.
Security scanned
Passed our 8-point scan before listing
Fresh listing
Recently published to Agensi
30-day refund
Not a fit? Get your money back
Trust & safety
Security scanned
Verified clean today
- Passed all security checks, Safe to install
Needs access to
Json Schema
Owasp
Github