versus
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS] (LOW): The skill makes automated network requests to
vibeindex.aito fetch resource metadata. This domain is not part of the pre-approved trusted sources list. - [DATA_EXFILTRATION] (LOW): The skill reads the local
package.jsonfile. While this is used to tailor recommendations to the user's project, accessing local filesystem metadata and potentially reflecting it in output that is influenced by external API data presents a minor exposure surface. - [PROMPT_INJECTION] (LOW): The skill is vulnerable to indirect prompt injection because it ingests untrusted data from an external API and uses it to generate the final response.
- Ingestion points: Resource descriptions and metadata from
https://vibeindex.ai/api/resources. - Boundary markers: None identified; external content is interpolated directly into the comparison prompt.
- Capability inventory: Reads local
package.jsonfile. - Sanitization: None identified; the agent is instructed to 'Extract' and 'Present' fields directly from the API response.
Audit Metadata