carrefour-api
Pass
Audited by Gen Agent Trust Hub on Nov 12, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill references the installation of vendor-owned packages to facilitate integration.
- Includes recommendations for the Python package
scavioand Node.js packagesscavioand@scavio/mcp-server. These resources are maintained by the skill author (scavio-ai) and represent standard SDK deployment.
- Includes recommendations for the Python package
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external content from Carrefour product listings and user-generated reviews.
- Ingestion points: Data retrieved from the
/search,/product, and/reviewsendpoints, including product descriptions, ingredient lists, and customer feedback (SKILL.md). - Boundary markers: Not explicitly defined in the raw data flow, but implied by the use of structured JSON.
- Capability inventory: The agent possesses network capabilities to communicate with
api.scavio.devfor data fetching (SKILL.md). - Sanitization: Instructions provide strong guardrails, directing the agent to quote external content verbatim, avoid rewriting it into data structures, and translate only in surrounding context, which mitigates the risk of the agent accidentally executing instructions embedded in grocery data (SKILL.md).
- Ingestion points: Data retrieved from the
Audit Metadata