nano-pdf
Pass
Audited by Gen Agent Trust Hub on May 19, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill requires the installation of the
nano-pdflibrary from PyPI, which is a well-known service for package distribution. - [COMMAND_EXECUTION]: The skill relies on executing the
nano-pdfCLI tool to perform PDF editing tasks, which involves processing shell commands with dynamic arguments. - [PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface as it interpolates natural language instructions directly into a CLI command without sanitization.
- Ingestion points: Untrusted data enters the agent context via the
<instruction>argument in thenano-pdf editcommand (SKILL.md). - Boundary markers: There are no explicit delimiters or instructions to the LLM to ignore potentially adversarial commands embedded in the PDF content or the instruction string.
- Capability inventory: The skill invokes a CLI tool (
nano-pdf) that can modify local file system contents (PDFs). - Sanitization: There is no evidence of input sanitization, escaping, or validation of the instruction string before it is passed to the shell.
Audit Metadata