stirling-pdf

Pass

Audited by Gen Agent Trust Hub on Nov 4, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
  • [SAFE]: The skill provides legitimate instructions and scripts for self-hosting a PDF utility, focusing on standard administrative tasks.
  • [COMMAND_EXECUTION]: Instructions include the use of shell commands for service management and API interaction.
  • Examples include docker compose up, curl for API requests, and openssl for secret generation.
  • Secret handling is performed securely using read -rs to prevent echoing sensitive input to the terminal.
  • [EXTERNAL_DOWNLOADS]: The skill references external resources from reputable sources.
  • Downloads the Stirling PDF container image from the official registry docker.stirlingpdf.com.
  • Recommends fetching OCR language data from the official Tesseract tessdata GitHub repositories.
  • [INDIRECT_PROMPT_INJECTION]: The skill is designed to process user-supplied PDF documents.
  • Ingestion points: Processes PDF files via fileInput fields in API calls and local file reads in Python scripts (e.g., /srv/scans/inbox).
  • Boundary markers: None explicitly defined in the prompt templates, though the tool is used for discrete processing tasks.
  • Capability inventory: Includes file system writes (saving processed PDFs), subprocess execution via curl and docker, and network requests via the Python requests library.
  • Sanitization: Not applicable for binary PDF processing; the agent's role is to facilitate the tool's execution rather than interpreting content for decision-making.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 4, 2026, 12:41 PM