minimax-pdf

Pass

Audited by Gen Agent Trust Hub on Jun 22, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill facilitates the download and installation of necessary libraries and assets from well-known services.
  • Fetches design fonts from Google Fonts via CSS @import rules in the cover HTML templates.
  • Automatically installs required Python and Node.js packages using standard package managers (pip, npm) during the setup or execution phase.
  • [COMMAND_EXECUTION]: The skill utilizes shell scripts and subprocesses to coordinate its multi-stage PDF production pipeline and manage dependencies.
  • The make.sh script executes various Python and Node.js components to transform content into a final PDF.
  • Multiple Python scripts, including fill_write.py, merge.py, and render_body.py, utilize subprocess.check_call to verify and install missing dependencies at runtime.
  • [PROMPT_INJECTION]: The skill's reformatting feature (Route C) creates a surface for indirect prompt injection by ingesting and parsing untrusted external data.
  • Ingestion points: The scripts/reformat_parse.py script reads and parses user-supplied Markdown, PDF, TXT, and JSON files from the project environment.
  • Boundary markers: No explicit instructions or delimiters are present to warn the agent about ignoring embedded instructions within processed documents.
  • Capability inventory: The skill can execute shell commands via make.sh, write to the local file system (PDFs and configuration files), and perform network operations via Playwright.
  • Sanitization: Basic regex-based sanitization is used to convert Markdown into ReportLab XML markup, primarily focusing on visual formatting rather than instruction filtering.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 22, 2026, 02:17 PM
Security Audit — agent-trust-hub — minimax-pdf