The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses a local validation script, scripts/validate-design-md.mjs, to verify the structure and data integrity of the generated reports. The script uses a secure implementation, spawning a Python subprocess to perform safe YAML parsing via yaml.safe_load, preventing arbitrary code execution during the validation phase.- [EXTERNAL_DOWNLOADS]: The skill references standard development tools including the pyyaml library from the Python Package Index and the @google/design.md linter from Google's official npm organization. These are recognized, well-known, and trusted sources for utility software.- [INDIRECT_PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface as it ingests and processes data from external public websites. The instructions mitigate this risk by requiring the agent to verify visual claims against screenshots and by enforcing a strict schema for the evidence sidecar, preventing the agent from being misled by adversarial content on analyzed pages.

awwwards-design-md