data-build-dashboard
Pass
Audited by Gen Agent Trust Hub on Apr 18, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill fetches the Chart.js library from the jsDelivr CDN, which is a well-known service for hosting web development resources. The implementation includes a subresource integrity hash to ensure the script has not been tampered with.
- [COMMAND_EXECUTION]: The instructions direct the agent to save the generated content as an HTML file and subsequently open it in the default web browser.
- [PROMPT_INJECTION]: The skill exhibits an attack surface for indirect prompt injection as it processes untrusted data from various sources (CSV files, pasted text, and data warehouse results) and embeds it into a generated HTML file. Ingestion points: Data warehouse results, uploaded files, and pasted data in SKILL.md. Boundary markers: No specific delimiters are defined to separate untrusted data from the dashboard's logic. Capability inventory: The skill can write local files and trigger browser opening. Sanitization: The instruction to 'clean the data' is noted, but the skill lacks specific directives for HTML escaping or implementing a Content Security Policy to prevent potential script execution from malicious data.
Audit Metadata