Security audit
Security checks for vulnerabilities and agentic risk
This OCR skill is coherent and disclosed as a third-party API integration, but users should treat submitted PDFs as documents sent to pdfapihub.com.
Install only if you are comfortable sending PDFs, document URLs, and extracted OCR results to pdfapihub.com under that provider's terms. Avoid confidential, regulated, or customer documents unless your organization has approved that service and understands its retention and privacy practices.
The skill description says it extracts text using Tesseract OCR, which strongly implies local/offline processing, but the example sends the PDF to a third-party service at pdfapihub.com. This creates a material trust-boundary mismatch: users may expose sensitive document contents externally without realizing the skill is cloud-based rather than local.
The example performs a POST to an external OCR API and includes a document URL, causing document data or access to document contents to be transmitted to a third party. For OCR workflows, documents often contain invoices, IDs, contracts, or other sensitive data, so undisclosed external transmission materially increases confidentiality and compliance risk.
The skill instructs users to provide a PDF via URL, base64, or file upload and to authenticate against a third-party endpoint, but it does not clearly warn that document contents will be transmitted to an external OCR service. This is dangerous because users may submit sensitive invoices, legal records, insurance claims, or medical documents without informed consent about off-platform data sharing.
The example usage explicitly sends document data to an external service at pdfapihub.com, creating a real data egress path. In the context of OCR for invoices, claims, and legal documents, this can expose sensitive content to a third party if users assume the skill operates locally or do not understand the trust boundary.
curl -X POST https://pdfapihub.com/api/v1/pdf/ocr/parse \
-H "CLIENT-API-KEY: your_api_key" \
-H "Content-Type: application/json" \
-d '{
The request body sets "lang": "eng", and the sample response repeats "language": "eng", which indicates the skill is fixed to a specific language. Under the language/locale policy, forcing a language without user opt-in or justification is a natural-language policy concern.
The skill sends PDF content to an external third-party API, but the manifest does not clearly warn users that uploaded files or referenced document URLs will leave the local environment. This creates a meaningful data exposure and privacy risk, especially if users submit confidential, regulated, or internal documents under the assumption processing is local.
This JSON file describes a request/response example but provides no information about when the skill should or should not activate. For manifest files, missing specificity around trigger scope can lead to unintended invocation because there is no explicit activation phrase, constraint, or exclusion guidance.
This is a manifest file, so vague-trigger review applies. The description states broad capabilities like extracting text from scanned PDFs and supporting multiple languages, but it does not define any explicit trigger phrases, scope limits, or negative examples that would help avoid unintended invocation from generic OCR-related requests.
No suspicious patterns detected.