The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses subprocess.run() to execute standard utility binaries for its core features. This includes calling the GitHub CLI (gh) to retrieve user issues, and using LibreOffice (soffice) and pdftoppm for document conversion tasks. These executions are parameterized and restricted to the skill's operational requirements.
[INDIRECT_PROMPT_INJECTION]: An indirect injection surface is present in scripts-html-artifact/render-github-issues.py, which fetches external data (issue titles and comments) from GitHub. While the script performs deterministic rendering into a local HTML file, the ingestion of untrusted content represents a potential vector if the agent later processes this content for decision-making. The risk is minimized by the script's focus on structured data visualization.
[DYNAMIC_EXECUTION]: The skill performs local dynamic imports using importlib.import_module() in scripts-distinctive-frontend-design/validate_design.py and scripts-html-artifact/validate-artifact.py. These calls specifically load the local css_slop_rules.py module to perform design audits, representing a safe use of dynamic loading for modular code organization.
[EXTERNAL_DOWNLOADS]: The skill documentation references well-known external services such as unpkg.com for Three.js module imports and mentions the requirement for the Playwright chromium binary. These are standard dependencies for web rendering and 3D development workflows and originate from trusted or well-known infrastructure.

frontend