Fail
Audited by Gen Agent Trust Hub on Oct 6, 2026
Risk Level: HIGHPRIVILEGE_ESCALATIONCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [PRIVILEGE_ESCALATION]: The skill instructs the agent to execute "sudo apt-get install -y poppler-utils" to install system dependencies. Using sudo allows the agent to acquire administrative permissions.
- [COMMAND_EXECUTION]: The skill defines a rendering command "pdftoppm -png $INPUT_PDF $OUTPUT_PREFIX" that interpolates variables directly into a shell execution context. If file names or paths are provided by an untrusted source, this could lead to command injection.
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process external PDF documents, which could contain malicious instructions designed to influence the agent's behavior. [1] Ingestion points: PDF content is read and reviewed as part of the primary workflow (SKILL.md). [2] Boundary markers: No explicit delimiters or instructions to ignore embedded commands are specified for the processed content. [3] Capability inventory: The skill uses shell commands (pdftoppm) and Python libraries for extraction and generation. [4] Sanitization: There is no evidence of sanitization or filtering of the extracted text before it is presented to the agent.
- [EXTERNAL_DOWNLOADS]: The skill installs several standard Python libraries (reportlab, pdfplumber, pypdf) and system tools (poppler-utils) to perform PDF operations.
Recommendations
- AI detected serious security threats
Audit Metadata