fin-kg-benchmark-report
Pass
Audited by Gen Agent Trust Hub on Jul 17, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill's HTML template (
assets/report-template.html) fetches the Chart.js visualization library from a well-known CDN (jsdelivr.net). - [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it ingests untrusted data from local YAML files and interpolates it into an HTML report template without comprehensive sanitization.
- Ingestion points: The
scripts/generate_report.pyscript usesPath.rglobto discover YAML files andyaml.safe_loadto parse their content. - Boundary markers: No explicit delimiters or boundary markers are used to isolate the external data from the HTML/JavaScript context.
- Capability inventory: The script has the capability to write the generated HTML report to the local filesystem.
- Sanitization: While
json.dumpsis used to serialize data into theREPORT_DATAvariable within a script tag, this does not prevent script tag breakout (e.g.,</script>) if the YAML content is maliciously crafted, creating a potential Cross-Site Scripting (XSS) vector when the report is viewed. - [REMOTE_CODE_EXECUTION]: The script generates a new HTML file (
benchmark-report.html) by injecting data into an existing template. This dynamically generated file contains executable JavaScript that processes the injected JSON data.
Audit Metadata