Security audit
Security checks for vulnerabilities and agentic risk
This home-design skill is coherent and purpose-aligned, but users should understand that the redesign itself happens on Roomvana and may involve uploading photos there.
Before installing, be comfortable with Roomvana links being generated and with any redesign photo upload happening on Roomvana's website under its own account and privacy terms. Avoid uploading images that reveal addresses, valuables, people, documents, or other sensitive details.
The skill instructs the agent to run a bundled script and access Roomvana's live catalog over the network, but it does not declare any tool scope or permissions boundary. That mismatch can lead to over-broad execution in environments that rely on manifest-declared capabilities, increasing the chance of unintended network access or script execution beyond what users expect.
The invocation guidance is broad enough to match many generic decorating or home-improvement requests, which can cause the skill to activate when the user did not intend to use a third-party handoff workflow. In this skill's context, unintended invocation matters because it may funnel users toward an external service and photo upload flow without sufficiently clear consent up front.
The skill directs users to upload photos of their home to a third-party website but does not provide a prominent privacy warning before that handoff. Photos of interiors and exteriors can expose sensitive information such as addresses, valuables, family details, and layout information, so users need clear notice that their data leaves the assistant's environment.
Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.
python3 studio_link.py catalog-md # regenerates references/catalog.md
Every id is checked against Roomvana's live catalog (GET
https://api.roomvana.co/options, public, no auth), so a link never names a room,
style or finish the studio can't render. The combination rules mirror what the
studio itself accepts from a URL (roomvana.ai/design reads ?room= &mode=
&surface= &style= once on load).
Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.
python3 studio_link.py catalog-md # regenerates references/catalog.md
Every id is checked against Roomvana's live catalog (GET
https://api.roomvana.co/options, public, no auth), so a link never names a room,
style or finish the studio can't render. The combination rules mirror what the
studio itself accepts from a URL (roomvana.ai/design reads ?room= &mode=
&surface= &style= once on load).
Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.
python3 studio_link.py catalog-md # regenerates references/catalog.md
Every id is checked against Roomvana's live catalog (GET
https://api.roomvana.co/options, public, no auth), so a link never names a room,
style or finish the studio can't render. The combination rules mirror what the
studio itself accepts from a URL (roomvana.ai/design reads ?room= &mode=
&surface= &style= once on load).
Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.
python3 studio_link.py catalog-md # regenerates references/catalog.md
Every id is checked against Roomvana's live catalog (GET
https://api.roomvana.co/options, public, no auth), so a link never names a room,
style or finish the studio can't render. The combination rules mirror what the
studio itself accepts from a URL (roomvana.ai/design reads ?room= &mode=
&surface= &style= once on load).
This code performs an HTTP request to a third-party service to validate options, but the operation has no confirmation prompt or user-facing notice at execution time. While the module docstring mentions the live catalog lookup, users invoking the script may not see that disclosure in normal CLI use.
No suspicious patterns detected.