epub-accessibility

Pass

Audited by Gen Agent Trust Hub on Sep 26, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONDYNAMIC_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted content from EPUB files, which can include XHTML and XML documents. A maliciously crafted EPUB could contain instructions designed to influence the agent's behavior during the audit process. * Ingestion points: Content documents (*.xhtml, .html) and package metadata (.opf) processed as described in references/epub-accessibility-rule-set.md. * Boundary markers: None identified; the agent is instructed to read and analyze the raw file contents directly. * Capability inventory: File system operations including directory creation, extraction, and reading via shell commands like cat and grep. * Sanitization: No explicit sanitization or escaping of external content is defined.
  • [COMMAND_EXECUTION]: The reference documentation references/epub-accessibility-rule-set.md provides shell command templates (Bash and PowerShell) for auditing tasks, including file extraction (unzip, Expand-Archive) and metadata reading (cat).
  • [DYNAMIC_EXECUTION]: The skill includes shell script logic in references/epub-accessibility-rule-set.md, such as a Bash while loop used to iterate through files and execute grep commands for structural analysis.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 26, 2026, 05:59 AM