stitched-full-page-capture

Pass

Audited by Gen Agent Trust Hub on Sep 21, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The script scripts/stitch_full_page_capture.mjs executes external binaries ffmpeg and sips using spawn and execFile. These tools are used for image processing and metadata extraction. It also launches a browser instance using Playwright for web automation.
  • [EXTERNAL_DOWNLOADS]: The script attempts to load the playwright library from the current working directory's Node.js environment.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes data from external sources, specifically manifest.json files and the content of web pages navigated to via item.pageUrl. Ingestion points: Reads manifest.json file paths and navigates to URLs provided in the manifest items. Boundary markers: No explicit boundary markers or 'ignore' instructions are used when processing manifest data. Capability inventory: The skill has the ability to write/overwrite files via ffmpeg and fs.writeFile, and to delete temporary directories via fs.rm. It also has network access via the Chromium browser. Sanitization: There is no path sanitization on the fullPageImage or sectionImages file paths defined in the manifest, which represents a potential surface for path traversal if an agent is tricked into processing a malicious manifest.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 21, 2026, 01:30 AM
Security Audit — agent-trust-hub — stitched-full-page-capture