citation-data-extractor
Pass
Audited by Gen Agent Trust Hub on Mar 25, 2026
Risk Level: SAFE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to process untrusted
html_contentfrom inspection reports. This creates a surface for indirect prompt injection if the source HTML contains malicious instructions. However, the skill does not utilize sensitive tools, perform network operations, or execute system commands, significantly limiting the impact of such an attack. - Ingestion points: Processes
html_contentin the extraction pipeline. - Boundary markers: No specific boundary markers or delimiters for the untrusted content are defined in the extraction strategy.
- Capability inventory: No high-risk capabilities (shell, network, file-write) are requested or used.
- Sanitization: The skill focuses on regex and pattern matching for extraction but does not explicitly define sanitization steps for the input HTML.
Audit Metadata