build-visualizations
Pass
Audited by Gen Agent Trust Hub on Nov 20, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill fetches data from external datasets using the
fetch_datasettool to populate visualizations and dynamic content blocks (MarkdownViz). This creates a surface where malicious instructions embedded in the data could attempt to influence the agent's output or behavior. \n - Ingestion points: Dataset fields and measures retrieved via
fetch_datasetas described inSKILL.md. \n - Boundary markers: The instructions specify the use of
@md ... ;;heredocs for multi-line prose and labels, which helps delimit data-driven content. \n - Capability inventory: The skill uses
fetch_dataset,generate_aql,generate_viz,execute_viz, andsearch_docsto build and validate visualizations. \n - Sanitization: The instructions explicitly forbid the use of JavaScript within
MarkdownVizblocks, stating they must be non-interactive, which reduces the risk of malicious script execution.
Audit Metadata