screenshot-docs

Pass

Audited by Gen Agent Trust Hub on Sep 21, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The script scripts/install_playwright_capture.sh facilitates the installation of the playwright package and Chromium browser via npm. These resources originate from well-known, trusted technology providers and registries.
  • [COMMAND_EXECUTION]: The utility scripts/capture_cli.sh executes arbitrary shell commands provided as arguments ("$@") to capture their output for rendering into visual documentation. This provides a functional but broad interface for shell command execution within the skill's workflow.
  • [DYNAMIC_EXECUTION]: Several scripts utilize process execution to interact with system and development tools:
  • scripts/screenshot_age.py executes git to retrieve file versioning metadata.
  • scripts/render_artifact_libreoffice.sh invokes soffice (LibreOffice) and magick (ImageMagick) to process and render document artifacts.
  • scripts/mini_capture_window.sh uses osascript (AppleScript) to control macOS application focus and retrieve window coordinates.
  • [INDIRECT_PROMPT_INJECTION]: The skill exhibits a vulnerability surface where untrusted data from a target repository (such as command outputs, source code, or web page content) could influence the execution of shell commands or browser automation.
  • Ingestion points: Target repository files (README, package.json, source code), CLI tool outputs, and web application content are read by the agent to determine capture parameters.
  • Boundary markers: The skill does not implement specific delimiters or instructions to sanitize application output or isolate it from the agent's decision-making logic.
  • Capability inventory: Capability includes shell command execution, filesystem writes (docs/screenshots/), and network requests via the Playwright browser.
  • Sanitization: No explicit sanitization is performed on data interpolated into command-line arguments for the capture scripts.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 21, 2026, 09:19 AM