research-report-reader

Pass

Audited by Gen Agent Trust Hub on Mar 30, 2026

Risk Level: SAFE
Full Analysis
  • [COMMAND_EXECUTION]: The skill utilizes the pdftotext command-line utility as a fallback method to extract text from user-uploaded PDF files. This is a legitimate functional requirement for a document analysis skill and does not involve arbitrary or unsafe command execution.
  • [PROMPT_INJECTION]: Indirect Prompt Injection Surface. The skill is designed to ingest and process content from external PDF files provided by the user. This creates a surface for indirect prompt injection, where malicious instructions embedded in a document could attempt to influence the agent's summary or translation. However, the skill lacks dangerous capabilities such as network access or high-privilege system commands, limiting the potential impact.
  • Ingestion points: Reads uploaded PDF files using the Read tool and pdftotext utility in Step 1.
  • Boundary markers: The instructions do not define specific delimiters or "ignore" instructions for the processed document content.
  • Capability inventory: Limited to file system discovery (ls), document reading, and writing Markdown files to the workspace (Edit).
  • Sanitization: No explicit sanitization of document content is implemented.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 30, 2026, 10:09 PM