skills/krishnanaicker/padhai/padhai/Gen Agent Trust Hub

padhai

Pass

Audited by Gen Agent Trust Hub on Nov 19, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill utilizes the subprocess.run function in scripts/build.py and scripts/qa.py to perform essential document processing tasks. These tasks include launching a headless Chromium-family browser to print HTML to PDF, running a local Node.js script (render_math.js) to pre-render KaTeX formulas, and optionally calling the pdftoppm utility to convert PDF pages to images for quality checks. These executions are scoped to local utilities and distributed scripts.
  • [EXTERNAL_DOWNLOADS]: In scripts/build.py, the skill is configured to use a public CDN (cdn.jsdelivr.net) to load the KaTeX library for math rendering if a local Node.js installation is not detected. This is a common and legitimate practice for web-based rendering tools.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes user-provided syllabus data and HTML answer parts. While this represents a data ingestion surface, the static analysis shows that the skill treats this data strictly as content for rendering and layout. It does not parse the content for instructions that would influence the agent's behavior beyond the scope of note generation.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 19, 2026, 06:37 AM