Insurance Claims Intelligence
PassAudited by VirusTotal on May 11, 2026.
Overview
Type: OpenClaw Skill Name: insurance-claims-intelligence Version: 2.0.0 The skill bundle is a comprehensive documentation and template set for insurance claims processing, including OCR integration guidance and liability determination frameworks. It contains Python code snippets (e.g., in references/claims_ocr_tech.md and references/claims_liability_engine.md) that are clearly marked as advisory examples, follow security best practices like using environment variables for API keys, and include extensive disclaimers regarding data privacy (PIPL compliance) and the necessity of human review.
Findings (0)
Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.
If you run the integration examples, your OCR or LLM account credentials may be used and could incur cost or expose account access if mishandled.
The skill expects user-supplied service credentials for optional OCR/LLM examples. This is expected for the stated purpose, but the user must control key scope and storage.
Requires API credentials: YES — you must provide your own OCR/LLM API keys
Use least-privilege provider keys, store them in environment variables or a secret manager, and revoke unused keys.
Medical invoice images and claimant details could be transmitted to a third-party OCR provider if the example is implemented.
The OCR example sends selected invoice images to a disclosed cloud OCR provider. Medical invoices are sensitive, but the document also instructs users to obtain consent, redact PII, and check retention terms.
with open(image_path, "rb") as f: ... img_b64 = base64.b64encode(f.read()) ... url = "https://aip.baidu.com/rest/2.0/ocr/v1/medical_invoice" ... resp = requests.post(url, data=payload, headers=headers)
Only upload user-approved documents, redact unnecessary personal information, verify provider retention/cross-border terms, and prefer private/on-prem OCR for production.
A user could mistakenly send a draft as an official approval, denial, payout, or fraud-related notice.
Some report templates use formal claim-status and payment language that could look like a final decision if copied directly, even though the broader artifacts require licensed human review.
**处理状态**: {{AUTO_PASS | MANUAL_REVIEW | HIGH_RISK | REJECT}} ... 本次理赔已通过智能审核，赔款将于3个工作日内转入您指定的银行账户。Keep all generated reports clearly labeled as drafts until a licensed claims professional reviews and approves them; consider replacing auto-pass wording with draft recommendation wording.