build-data-app
Pass
Audited by Gen Agent Trust Hub on Oct 28, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill fetches external libraries including ECharts, Basecoat UI, and Inter fonts from well-known content delivery networks such as jsDelivr, Cloudflare, and Google Fonts. The instructions explicitly require the use of pinned versions and integrity hashes (SHA-384) to prevent supply chain attacks.
- [COMMAND_EXECUTION]: The skill utilizes a set of local vendor-provided CLI tools (e.g.,
anfra show,anfra query,anfra validate) to manage datasets, perform data queries, and verify the status of the local serving environment. - [INDIRECT_PROMPT_INJECTION]: The skill processes dataset metadata such as labels and descriptions to inform the design of the dashboard. While this represents a potential ingestion point for external data, the risk is mitigated by the skill's instruction to use structured SDK calls and specific query syntax rather than raw string interpolation.
- [DYNAMIC_EXECUTION]: The primary function of this skill is to generate self-contained HTML files that include JavaScript for data visualization and user interaction. These scripts operate within a sandboxed frame defined by the Anfra environment.
Audit Metadata