paper-siphon

Pass

Audited by Gen Agent Trust Hub on Apr 8, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill uses the uvx tool to fetch the paper-siphon package from the Python Package Index (PyPI). This is a legitimate dependency provided by the skill author for PDF processing.
  • [REMOTE_CODE_EXECUTION]: By using uvx, the skill downloads and executes the paper-siphon package on the host system. This is the intended behavior for running the conversion utility without a permanent installation.
  • [COMMAND_EXECUTION]: The skill documentation provides shell commands that the agent is expected to execute to process PDF inputs. These commands take local file paths or external URLs as arguments.
  • [PROMPT_INJECTION]: The skill creates a surface for indirect prompt injection by converting external, untrusted PDF documents into text that the agent then processes. 1. Ingestion points: External PDF files and URLs. 2. Boundary markers: None specified in the instructions to delimit extracted text. 3. Capability inventory: The skill possesses file reading, network fetching, and file writing capabilities. 4. Sanitization: No sanitization of the extracted content is mentioned.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 8, 2026, 05:36 PM
Security Audit — agent-trust-hub — paper-siphon