grocery-cart
Pass
Audited by Gen Agent Trust Hub on Aug 15, 2026
Risk Level: SAFE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill handles data from external recipes and store APIs which could theoretically contain malicious instructions. The skill proactively mitigates this risk by instructing the agent to treat all external data as untrusted and requiring a "confirmation gate" where the user must approve the final shopping list before any cart actions are taken.
- Ingestion points: External recipe content and API/tool responses (SKILL.md).
- Boundary markers: Explicitly instructs the agent that external data is untrusted and should not be treated as instructions.
- Capability inventory: Perform network requests to store APIs, store user preferences, and update kitchen inventory (SKILL.md).
- Sanitization: Implements a mandatory user confirmation step before performing any cart additions.
- [EXTERNAL_DOWNLOADS]: The skill connects to api.peristyle.io and mcp.peristyle.io to fetch recipe data, match products, and manage user preferences. These domains are owned by the skill's author and are necessary for the skill's core functionality.
- [COMMAND_EXECUTION]: The documentation provides installation commands for the user to set up the MCP server (e.g., npx skills add, claude mcp add). These are instructions for the user to execute in their local environment and are not automated actions performed by the agent.
Audit Metadata