Security audit
Security checks for vulnerabilities and agentic risk
This translation skill is coherent, but it directs the agent to send emails, text, documents, and website/CMS content to an external service without enough user-facing consent and data-handling guidance.
Review this skill before installing if you may translate private, regulated, proprietary, unpublished, or account-related content. Use it only if you are comfortable sending that material, and your email address for signup, to TranslateFlow/VCG; avoid submitting secrets or sensitive documents unless you have approval and understand the provider's data handling.
The manifest description is excessively broad and packed with generic translation-related trigger terms, which can cause the skill to activate for many loosely related requests. Overbroad routing increases the chance that users are funneled into workflows that collect personal data or transmit content to a third-party API without the user specifically intending to use this external service.
The skill instructs the agent to collect the user's email and submit it to an external API, but it does not clearly warn the user that their email will be shared with a third party. This creates a privacy and consent problem because users may believe they are only interacting locally with the agent rather than registering with an external service.
This step performs direct external transmission of the user's email address to a third-party endpoint to create an API key. External transmission itself is expected for the feature, but without a consent/privacy notice and clear trust boundary, it becomes a real privacy vulnerability because personally identifiable information is sent off-platform by the agent.
curl -X POST https://translateflow.vosscg.com/v1/keys \
-H "Content-Type: application/json" \
-d '{"email":"user@example.com"}'
The translation examples send arbitrary user text to an external API without warning that the content leaves the local environment. Users may submit sensitive business, personal, medical, or legal text assuming the assistant is handling it privately, resulting in unintended data disclosure to the third-party service.
The document translation workflow uploads whole files to a third-party service but does not warn users that document contents and metadata will be externally transmitted. Because documents commonly contain confidential, regulated, or proprietary information, silent upload to an outside processor materially increases privacy, compliance, and data leakage risk.
The website and CMS localization examples send site URLs, page content, metadata, and CMS text to an external service without an explicit warning. This can expose unpublished content, internal URLs, proprietary copy, SEO data, or customer-facing material to a third party when the user may not realize external processing is occurring.
Detected: suspicious.exposed_secret_literal